WO2010004744A1 - 経路危険性評価装置 - Google Patents
経路危険性評価装置 Download PDFInfo
- Publication number
- WO2010004744A1 WO2010004744A1 PCT/JP2009/003183 JP2009003183W WO2010004744A1 WO 2010004744 A1 WO2010004744 A1 WO 2010004744A1 JP 2009003183 W JP2009003183 W JP 2009003183W WO 2010004744 A1 WO2010004744 A1 WO 2010004744A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- evaluation
- route
- risk
- unit
- damage
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1656—Programme controls characterised by programming, planning systems for manipulators
- B25J9/1664—Programme controls characterised by programming, planning systems for manipulators characterised by motion, path, trajectory planning
- B25J9/1666—Avoiding collision or forbidden zones
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0214—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with safety or protection criteria, e.g. avoiding hazardous areas
Definitions
- the present invention relates to a route risk evaluation apparatus for planning a movement action of a mobile device based on the risk of the movement route of a mobile device such as a robot.
- Patent Document 1 a robot that moves while holding an object is known (for example, see Patent Document 1).
- the robot grips an object, it is necessary to accurately grasp the relative position between the robot and the object to be gripped.
- the gripping may fail because the position measurement accuracy cannot be obtained.
- a gripping target object that collides with the robot may be bounced by the robot and move to another location. Therefore, in order to prevent the gripping target object from being bounced and moved to another location, a method disclosed in Patent Document 2 has been proposed. In this method, the robot is moved at high speed to the vicinity of the object to be grasped.
- the robot After approaching a predetermined distance to the object to be grasped, the robot is made to approach the object to be grasped at a low speed.
- a method of adding a position sensor to a robot for example, see Patent Document 3 has been proposed. If such a method is used, even if the robot may come into contact with the object to be grasped, the speed at the time of collision is low. For this reason, it is less likely that the gripping target object is bounced and moves to another location.
- the robot used in the home environment has a lower object position recognition accuracy than the industrial robot, and therefore, the collision between the robot and the object occurs.
- the position where the position sensor can be installed is limited to a range that does not interfere with humans.
- Various types of objects are conceivable as the gripping target object and the objects around the gripping target object. Therefore, it is easy to misrecognize the position of the object to be gripped, and collision between the robot and the object is likely to occur.
- the second case is when the robot cannot be controlled accurately compared to an industrial robot and a collision between the robot and an object occurs.
- the floor on which the robot operates is more diverse than in a factory, so the robot may take an unstable posture.
- the robot itself may vibrate due to the vibration of a motor or the like equipped to cool the robot. Therefore, the control of the robot tends to be inaccurate, and the collision between the robot and the object tends to occur.
- the third is a case where a collision between a robot and an object occurs due to severe restrictions on task execution time.
- a robot used in a home environment is required to move at a certain high speed so as not to interfere with humans. Therefore, when an action that prioritizes moving at a high speed is planned rather than avoiding a collision, a collision between the robot and an object is likely to occur.
- the robot may collide with the object. Therefore, the robot or a part thereof may collide with a certain first object, and the first object or a part thereof may move due to the impact. Furthermore, the moved first object or a part thereof may contact or approach the surrounding second object. At that time, depending on the combination of the types of the first object and the second object, there is a problem that damage occurs due to contact or approach between the first object and the second object.
- the first example of this problem occurs when the robot moves. That is, when the robot is moving, it may collide with the first object existing on the route. Then, the collided first object moves and further contacts with or approaches a second object in the vicinity thereof to cause damage.
- the second example of this problem occurs when the robot grips an object. That is, when the robot grasps the object to be grasped, the robot may not be able to grasp the object to be grasped and may play it. Then, damage is caused when the flipped object to be gripped contacts or approaches the surrounding objects.
- the third example of this problem occurs when a robot carries an object. That is, when the robot transports an object, the first object to be transported may contact or approach the second object in the vicinity of the path for transporting the first object. Even in that case, depending on the combination of the types of the first object and the second object, damage occurs due to contact or approach.
- the present invention has been made in order to solve the above-described problems, and the damage caused by the movement of the robot due to the movement of the robot and the movement of the moved object in contact with or approaching the surrounding objects, or the risk thereof. It is an object of the present invention to provide a route risk evaluation device that can be reduced.
- a route risk evaluation device is a route risk evaluation device that evaluates the risk of a moving route of a mobile device, and is a movement that is information indicating the moving route of the mobile device.
- a path holding unit that holds path information, an object position attribute holding unit that holds position information indicating the positions of a plurality of objects existing in an environment in which the mobile device moves, and objects that exist in the environment are in contact with each other.
- a risk attribute combination information holding unit that holds information indicating the degree of damage when approaching, the movement route of the mobile device indicated by the movement route information held in the route holding unit, and the object position attribute holding unit A first object that may be moved by contacting or approaching the mobile device based on the relationship with the positions of the plurality of objects indicated by the position information held in
- the secondary collision possibility evaluation unit that evaluates the possibility of contacting or approaching two objects
- a risk attribute combination evaluation unit that determines the degree of damage when the second object comes into contact with or approaches, the evaluation result by the secondary collision possibility evaluation unit, and the damage level determined by the risk attribute combination evaluation unit
- a route evaluation unit that evaluates the risk of the travel route with respect to the mobile device.
- the possibility of a secondary collision in which the first object and the second object come into contact or approach, and the secondary The risk of the movement route for the mobile device can be evaluated from both the degree of damage when the collision occurs. As a result, it is not necessary to select a movement route that has a high damage risk due to the secondary collision as the movement route of the mobile device, and the damage risk due to the secondary collision can be reduced.
- the present invention can be realized not only as a route risk evaluation apparatus including such a characteristic processing unit, but also as a route risk having a characteristic processing unit included in the route risk evaluation apparatus as a step. It can also be realized as an evaluation method, or as a program for causing a computer to execute characteristic steps included in the route risk evaluation method. Needless to say, such a program can be distributed via a computer-readable recording medium such as a CD-ROM (Compact Disc-Read Memory) or a communication network such as the Internet.
- a computer-readable recording medium such as a CD-ROM (Compact Disc-Read Memory) or a communication network such as the Internet.
- the present invention it is possible to reduce the damage caused by the movement of the robot due to the movement of the robot, and the damage caused by the moved object coming into contact with or approaching the surrounding object or its risk.
- FIG. 1 is a block diagram showing the configuration of the action planning apparatus according to the first embodiment.
- FIG. 2 is a diagram illustrating an example of object position attribute information held by the object position attribute holding unit according to the first embodiment.
- FIG. 3 is a diagram illustrating an example of the arrangement of objects in the first embodiment.
- FIG. 4 is a diagram illustrating an example of a route held by the route holding unit according to the first embodiment.
- FIG. 5 is a diagram illustrating an example of an angle formed by a straight line connecting a gripping target object and a peripheral object and a robot path in the first embodiment.
- FIG. 6 is a diagram illustrating an example of the risk attribute combination held by the risk attribute combination information holding unit according to the first embodiment.
- FIG. 1 is a block diagram showing the configuration of the action planning apparatus according to the first embodiment.
- FIG. 2 is a diagram illustrating an example of object position attribute information held by the object position attribute holding unit according to the first embodiment.
- FIG. 3 is a diagram illustrating an example of the
- FIG. 7 is a flowchart showing a flow of processing executed by the behavior planning apparatus in the first embodiment.
- FIG. 8 is a diagram for explaining a method of calculating the secondary collision possibility direction evaluation value and the secondary collision possibility distance evaluation value in consideration of the size of the surrounding object.
- FIG. 9 is a diagram for explaining a method of calculating the secondary collision possibility direction evaluation value and the secondary collision possibility distance evaluation value in consideration of the size of the surrounding object.
- FIG. 10 is a diagram for explaining a method of calculating the secondary collision possibility direction evaluation value and the secondary collision possibility distance evaluation value in consideration of the sizes of the gripping target object and the surrounding objects.
- FIG. 11 is a diagram illustrating a configuration of the robot system according to the second embodiment.
- FIG. 12 is a diagram illustrating a configuration of a robot arm according to the third embodiment.
- FIG. 13 is a block diagram illustrating a configuration of the action planning apparatus according to the third embodiment.
- FIG. 14 is a block diagram illustrating a configuration of the behavior planning apparatus according to the fourth embodiment.
- FIG. 15 is a flowchart showing a flow of processing executed by the behavior planning apparatus in the fourth embodiment.
- FIG. 16 is a diagram illustrating an initial path of the robot according to the fifth embodiment.
- FIG. 17 is a diagram showing an improved route of the robot in the fifth embodiment.
- FIG. 18 is a block diagram showing a configuration of the behavior planning apparatus in the fifth embodiment.
- FIG. 19 is a diagram illustrating an example of a plurality of paths that can execute the same task in the sixth embodiment.
- FIG. 19 is a diagram illustrating an example of a plurality of paths that can execute the same task in the sixth embodiment.
- FIG. 20 is a block diagram illustrating a configuration of the behavior planning apparatus according to the sixth embodiment.
- FIG. 21 is a block diagram illustrating a configuration of the behavior planning apparatus according to the seventh embodiment.
- FIG. 22 is a diagram illustrating an example of primary collision risk attribute information held by the primary collision risk attribute information holding unit according to the seventh embodiment.
- FIG. 23 is a flowchart showing a flow of processing executed by the behavior planning apparatus in the seventh embodiment.
- FIG. 24 is a block diagram illustrating a configuration of the action planning apparatus according to the eighth embodiment.
- FIG. 25 is a flowchart showing a flow of processing executed by the behavior planning apparatus in the eighth embodiment.
- an action planning apparatus that plans actions of a robot that holds an instructed object will be described.
- the behavior planning apparatus determines a moving speed when the robot is brought close to the gripping target object when the robot grips the gripping target object.
- ⁇ About secondary collision> When the robot approaches the gripping target object, the robot may not grip the gripping target object and may collide with the gripping target object and flip the gripping target object (hereinafter referred to as “primary collision”). .
- the cause is due to the inaccuracy of the position of the acquired object, the inaccuracy of the control of the robot, or the high-speed movement of the robot in accordance with the demand for high-speed movement.
- an object bounced by the robot may further contact or approach the surrounding object (hereinafter referred to as “secondary collision”). At that time, damage may occur depending on the combination of the first object in direct contact with the robot and the second object in contact with the first object.
- FIG. 1 is a block diagram illustrating a configuration of an action planning apparatus 100 that plans actions of a robot that holds an instructed object.
- the action planning apparatus 100 includes a robot position acquisition unit 101, a starting point position holding unit 102, an object position attribute acquisition unit 103, an object position attribute holding unit 104, a gripping target information acquisition unit 105, a route generation unit 106, a route holding unit 107, 2 Next collision possibility direction evaluation unit 108, secondary collision possibility distance evaluation unit 109, danger attribute combination information holding unit 111, danger attribute combination evaluation unit 112, route evaluation unit 113, evaluation holding unit 114, action planning unit 115, and plan A holding unit 116 is included.
- the action planning apparatus 100 is realized by executing a program for realizing the function of each processing unit constituting the action planning apparatus 100 on the CPU of the computer.
- the robot position acquisition unit 101 is a processing unit that acquires the position of the robot and stores the acquired robot position in the start position holding unit 102 as the start point of the movement path of the robot.
- the position of the robot is acquired by, for example, GPS (Global Positioning System), or acquired by performing image processing on an image obtained from an environment in which the robot moves or a camera installed in the robot. To do.
- GPS Global Positioning System
- the environment in which the robot moves is referred to as “environment”
- the environment in which the robot moves or the camera installed in the robot is referred to as “camera”.
- the starting point position holding unit 102 is a storage device that stores the position of the robot that is the starting point of the movement path of the robot acquired by the robot position acquiring unit 101, and includes, for example, an HDD (Hard Disk Drive).
- HDD Hard Disk Drive
- the object position attribute acquisition unit 103 is a processing unit that acquires the position and attribute information of an object existing in the environment.
- the attribute information is information indicating the type of object such as “cake”, “garlic”, “magnet”, and “flexible disk”.
- an RFID (Radio Frequency IDentification) tag is used to acquire the object position and attribute information. That is, an RFID tag in which attribute information is written is assigned in advance to each object in the environment, and the position and attribute information of the object are acquired by a known RFID wireless technology. Also, for example, image processing is used. That is, the position and attribute information is acquired by applying a known image processing technique to the image information obtained from the camera.
- RFID Radio Frequency IDentification
- the object position attribute holding unit 104 is a storage device that holds the position and attribute information of each object acquired by the object position attribute acquiring unit 103.
- FIG. 2 is a diagram illustrating an example of the object position attribute information held by the object position attribute holding unit 104.
- the object position attribute holding unit 104 holds zero or more pieces of object position attribute information composed of a set of information indicating the object position and object attribute information. In the example of FIG. 2, it is shown that there is a cake at the three-dimensional position (50, 15, 0), garlic at (40, 35, 0), and a magnet at (30, 45, 0). .
- FIG. 1 is a storage device that holds the position and attribute information of each object acquired by the object position attribute acquiring unit 103.
- FIG. 2 is a diagram illustrating an example of the object position attribute information held by the object position attribute holding unit 104.
- the object position attribute holding unit 104 holds zero or more pieces of object position attribute information composed of a set of information indicating the object position and object attribute information.
- FIG. 3 is a diagram illustrating the positional relationship of objects when the cake 402, the garlic 403, and the magnet 404, which are objects indicated by the object attribute information, are arranged at the object positions shown in this example.
- FIG. 3 shows an example in which the robot 401 exists at the three-dimensional position (10, 35, 0) in addition to the cake 402, the garlic 403, and the magnet 404.
- the grip target information acquisition unit 105 is a processing unit that acquires the position and type of the grip target object. Specifically, for example, the gripping target information acquisition unit 105 acquires the type of gripping target object voice uttered by the user using a voice input device (not shown). The gripping target information acquisition unit 105 acquires the type of the gripping target object by performing voice analysis on the acquired voice using a known voice analysis process. The grip target information acquisition unit 105 acquires the position of the grip target object having the same name as the acquired type from the object position attribute holding unit 104.
- the grip target information acquisition unit 105 outputs the acquired position of the grip target object to the secondary collision possibility direction evaluation unit 108, the secondary collision possibility distance evaluation unit 109, and the route evaluation unit 113, and acquires the acquired attribute information.
- the risk attribute combination evaluation unit 112 outputs the result.
- the method of specifying the object to be grasped is not limited to this example, and the position and attribute information of the object held by the object position attribute holding unit 104 is displayed on a display device (not shown), and one of them is held by the user's selection. It may be determined as a target object. Further, the gripping target object may be determined from the user's behavior and other information acquired from the outside. For example, the behavior planning apparatus obtains a forecast from a processing unit that obtains a weather forecast (not shown), and determines when the user wears shoes when the forecast is raining or on a day when the user wears shoes. It doesn't matter.
- the path generation unit 106 generates a path for moving the robot at the robot position that is the starting point of the movement path of the robot held by the start point position holding unit 102 to the position where the gripping target object acquired by the gripping target information acquisition unit 105 is located.
- a processing unit Specifically, for example, a directed line segment starting from the robot position and ending at a certain position of the object to be grasped is defined as a route.
- the route holding unit 107 is a storage device that holds the route generated by the route generation unit 106, and includes, for example, an HDD.
- FIG. 4 is a diagram illustrating an example of a route held by the route holding unit 107.
- the robot position that is the starting point of the movement path of the robot held by the starting point position holding unit 102 is (10, 35, 0)
- the position of the gripping target object is (40, 30, 0).
- the route is shown. This shows a path when the robot moves to hold garlic in the situation where an object (cake 402, garlic 403, and magnet 404) and the robot 401 exist in the positional relationship of FIG. .
- the secondary collision possibility direction evaluation unit 108 is a processing unit that evaluates how easily a specified object among objects existing in the environment is in a collision direction with respect to the object to be grasped. In the following description, this evaluation is called “secondary collision possibility direction evaluation”, and the evaluation value calculated at that time is called “secondary collision possibility direction evaluation value”.
- This secondary collision possibility direction evaluation is one of evaluations for estimating the possibility of a secondary collision. Note that the object may be specified by selecting arbitrary objects one by one from objects other than the gripping target object.
- the viewpoint of estimating the possibility of secondary collision by evaluating the direction of possibility of secondary collision relates to the angle formed by the straight line connecting the object to be grasped and surrounding objects and the robot path.
- the object to be grasped includes a direction that is likely to roll and a direction that is difficult to roll.
- the robot 401 moves in the X-axis direction 405.
- the garlic 403 is easy to roll in the positive direction 407 in the X-axis direction and is difficult to roll in the negative direction 406 in the X-axis direction.
- the garlic 403 is easy to roll in the same direction as the movement path of the robot 401, and is difficult to roll in the opposite direction. This coincides with the common sense that the object to be grasped hardly moves by pushing the robot itself in the direction blocked by the robot. Precisely, depending on conditions such as the presence of a wall in the same direction as the movement path of the robot or the unevenness of the floor, it may not be easy to roll in the same direction as the robot path. However, for the sake of simplification, condition determination that considers floor and wall conditions is ignored here.
- the possibility of collision is high when the angle formed between the straight line connecting the gripping target object and the surrounding object and the path of the robot is small, and the possibility of collision is low when the angle formed is large.
- the cake 402 at the position (50, 15, 0) has a high possibility of a secondary collision with the garlic 403 that is the object to be gripped, and the magnet 404 at the position (30, 45, 0) collides. The possibility of is low.
- the angle formed by the vector connecting the position of the object to be grasped and the position of the designated object and the path of the robot is ⁇ , and (1 + cos ⁇ ) / 2 is the secondary collision likelihood direction evaluation value.
- the maximum value of the secondary collision possibility direction evaluation value is 1, and the minimum value is 0.
- the value of the secondary collision possibility direction evaluation value is closer to the maximum value 1 as the specified object is in a direction in which a secondary collision is more likely to occur on the object to be grasped (as the formed angle ⁇ approaches 0 degrees). It becomes.
- the value of the secondary collision possibility direction evaluation value becomes closer to 0, which is the minimum value, as the direction is less likely to cause a secondary collision (as the formed angle ⁇ approaches 180 degrees). It is assumed that the formed angle ⁇ takes a value between 0 degrees and 180 degrees.
- FIG. 5 is an example showing an angle formed by a vector connecting the position of the object to be grasped and the position of the object in the environment and the path of the robot.
- ⁇ in the figure indicates the angle formed when the object to be gripped is garlic and the object in the environment is a cake. The angle formed by the above is shown.
- the secondary collision possibility direction evaluation unit 108 acquires the robot path from the path holding unit 107. That is, the start point and end point of the robot path are acquired.
- the secondary collision possibility direction evaluation unit 108 acquires the position of the designated object from the object position attribute holding unit 104.
- the secondary collision possibility direction evaluation unit 108 calculates the above-described secondary collision possibility direction evaluation value and outputs the calculation result to the route evaluation unit 113.
- the secondary collision possibility distance evaluation unit 109 is a processing unit that evaluates how easy the specified object is to collide with the object to be grasped. In the following description, this evaluation is called “secondary collision possibility distance evaluation”, and the evaluation value calculated at that time is called “secondary collision possibility distance evaluation value”.
- the secondary collision possibility distance evaluation is one of evaluations for estimating the possibility of a secondary collision.
- the viewpoint of estimation based on the secondary collision possibility distance evaluation relates to the distance between the object to be grasped and the surrounding object.
- the gripping target object collides with the robot and moves, the gripping target object eventually stops due to friction with the floor. Therefore, a peripheral object that is closer to the gripping target object is more likely to collide with the gripping target object. Further, even if the floor is free of friction and the gripping target object is slippery as far as possible, the more the peripheral object is away from the gripping target object, the less likely the peripheral object is in the direction of rolling of the gripping target object. The next collision is unlikely to occur.
- the object to be grasped is a bag containing soybeans
- the bag may collide with the robot and the soybeans may roll in all directions. If the number of soybeans is large enough and there is no friction on the floor, the importance of distance and direction is small. Originally, it may be possible to determine whether or not this is the case, but such determination processing is omitted in the first embodiment.
- the distance between the position of the object to be grasped and the position of the designated object is r
- the Euler number e to the ⁇ r power (e ⁇ r ) is the secondary collision possibility distance evaluation value.
- the secondary collision possibility distance evaluation value is the maximum value 1
- the secondary collision possibility distance evaluation value is the minimum value 0.
- the Euler number is also called a Napier number
- a specific value is 2.71828.
- the expression for evaluating the secondary collision possibility distance may be another expression as long as the value decreases as the distance increases. Further, it may be an equation according to a physical model considering energy loss due to friction.
- the secondary collision possibility distance evaluation unit 109 acquires the position of the designated object from the object position attribute holding unit 104.
- the secondary collision possibility distance evaluation unit 109 acquires the position of the gripping target object from the gripping target information acquisition unit 105.
- the secondary collision possibility distance evaluation unit 109 calculates a distance between two objects, calculates a secondary collision possibility distance evaluation value from the distance, and outputs the calculation result to the route evaluation unit 113.
- the dangerous attribute combination information holding unit 111 is information indicating what kind of attribute information object and what kind of attribute information object may cause damage when a secondary collision occurs.
- FIG. 6 is a diagram illustrating an example of danger attribute combination information.
- the combination of objects that cause damage in the case of a secondary collision may be a combination of garlic and cake, or a combination of a magnet and a flexible disk.
- the dangerous attribute combination evaluation unit 112 is a processing unit that evaluates whether a specified object is an object that may cause damage when a secondary collision with a gripping target object occurs. In the following description, this evaluation is called “danger attribute combination evaluation”, and the evaluation value calculated at that time is called “danger attribute combination evaluation value”.
- Danger attribute combination evaluation is an evaluation to estimate the degree of damage when a secondary collision occurs. Even if the object to be gripped and the surrounding object collide secondarily, damage does not always occur. For example, even if a clean spoon and fork collide secondarily, no damage will occur. Even if damage occurs, the degree of damage is different. For example, when a magnet and a flexible disk collide, the value of the data stored on the flexible disk becomes the amount of damage, which is sometimes serious damage. Conversely, as a case where the damage is small, a case where the magnet and the strawberry collide with each other can be considered. The time and energy for washing the strawberry or the time and energy for the robot to wash the strawberry will be damaged, but the degree of damage is small.
- the degree of damage varies, but in the first embodiment, the degree of damage is evaluated in two stages. That is, it is divided into a case where the degree of damage is so small that it can be ignored and a case where the damage cannot be ignored. If the degree of damage is small, a robot action plan that does not avoid secondary collision is performed. Whether the damage level is small or large is determined by determining whether the combination of the gripping target object and the surrounding object is a combination that causes damage.
- the dangerous attribute combination evaluation unit 112 performs a dangerous attribute combination evaluation on the specified object.
- the dangerous attribute combination evaluation unit 112 acquires the attribute information of the object to be grasped and the attribute information of the designated object.
- the dangerous attribute combination evaluation unit 112 determines whether the acquired combination of two pieces of attribute information is included in the dangerous attribute combination held by the dangerous attribute combination information holding unit 111.
- the risk attribute combination evaluation unit 112 outputs the determination result to the route evaluation unit 113.
- the risk attribute combination evaluation unit 112 sets the determination result to “1” when the combination of the two attribute information is included in the risk attribute combination, and sets the determination result to “0” when the combination is not included.
- the risk attribute combination information illustrated in FIG. 6 may include a monetary amount column so that the damage amount when the two attribute information objects collide is included.
- the risk attribute combination evaluation unit 112 uses the damage amount as the determination result, and if not included, the risk attribute combination evaluation unit 112 sets “0” as the determination result. Also good.
- the route evaluation unit 113 is a processing unit that evaluates the damage risk due to the secondary collision when the robot moves on the route held by the route holding unit 107.
- Embodiment 1 provides a robot behavior planning device that avoids a primary collision itself, which is a collision between a robot and an object to be grasped, but avoids a primary collision when the risk of damage is high.
- the “damage risk” is an evaluation with respect to a combination of the possibility that a secondary collision will occur and the degree of damage when a secondary collision occurs.
- Parameters for the simulation include those related to the object to be gripped, those related to the object around the object to be gripped, those related to the environment, and those related to the action plan of the robot.
- the parameters related to the gripping target object may include the size, shape, mass, friction coefficient, and position of the gripping target object.
- the parameters related to the surrounding objects include the size, shape, mass, coefficient of friction, and position of the surrounding objects.
- Environmental parameters may include floor shape, coefficient of friction, wind power, and atmospheric pressure.
- the robot action plan parameters regarding what position the robot moves at what speed can be considered.
- the secondary collision is evaluated by the secondary collision possibility direction evaluation evaluated by the secondary collision possibility direction evaluation unit 108 and the secondary collision possibility distance evaluation evaluated by the secondary collision possibility distance evaluation unit 109.
- the path evaluation unit 113 evaluates the damage risk by secondary collision for each candidate with the object to be grasped and the object in the environment other than the robot as the candidates for the secondary collision, and the total is calculated as the robot path. It is assumed that
- the damage risk for each candidate is calculated as the product of the possibility of a secondary collision and the degree of damage when a secondary collision occurs.
- the possibility of the secondary collision is evaluated based on the secondary collision possibility direction evaluation and the secondary collision possibility distance evaluation acquired from the secondary collision possibility direction evaluation unit 108 and the secondary collision possibility distance evaluation unit 109, respectively. Is done.
- the product of the secondary collision possibility direction evaluation and the secondary collision possibility distance evaluation is used as the evaluation value of the possibility of the secondary collision.
- the risk attribute combination evaluation acquired from the risk attribute combination evaluation unit 112 is used.
- the path evaluation unit 113 acquires the position of the gripping target object from the gripping target information acquisition unit 105.
- the path evaluation unit 113 designates each object (other than the gripping target object) corresponding to the attribute information held in the object position attribute holding unit 104, and performs the following processes (A) to (D).
- (A) Secondary collision possibility direction evaluation unit 108 obtains a secondary collision possibility direction evaluation.
- (B) Secondary collision possibility distance evaluation unit 109 acquires secondary collision possibility distance evaluation.
- the risk attribute combination evaluation unit 112 acquires the risk attribute combination evaluation.
- D Assume that (secondary collision possibility direction evaluation) ⁇ (secondary collision possibility distance evaluation) ⁇ (danger attribute combination evaluation) is the damage risk evaluation of the designated object.
- the route evaluation unit 113 calculates a total value of damage risk when each object calculated in the above (D) collides with the gripping target object, and the evaluation holding unit uses the calculated total value as a route evaluation value. 114.
- the evaluation holding unit 114 is a storage device that holds the evaluation of the route evaluated by the route evaluation unit 113, and includes, for example, an HDD.
- the behavior planning unit 115 is a processing unit that plans an action to avoid a secondary collision when the damage risk is high based on the route evaluation that is the damage risk of the route held by the evaluation holding unit 114. Specifically, if the route evaluation is larger than a predetermined value, an action that moves at a low speed is planned, and if it is smaller, an action that moves at a high speed is planned. Note that the predetermined value may not be set properly from the beginning, or may not match the situation on the way. Therefore, the predetermined value may be learned using a generally known learning algorithm and updated.
- the plan holding unit 116 is a storage device that holds information indicating an action plan performed by the action planning unit 115, and is configured by, for example, an HDD. Specifically, the plan holding unit 116 holds the speed of moving the route.
- FIG. 7 shows a flow of action plan processing performed by the action plan apparatus 100.
- the robot position acquisition unit 101 acquires the robot position and stores it in the start point position holding unit 102 (step S701). Specifically, when the robot 401 exists at the position shown in FIG. 3, the robot position acquisition unit 101 stores (10, 35, 0), which is the position of the robot 401, in the start point position holding unit 102.
- the object position attribute acquisition unit 103 acquires the object position attribute information of the object existing in the environment and stores it in the object position attribute holding unit 104 (step S702).
- the object position attribute information of “magnet”, “garlic”, and “cake” is stored in the object position attribute holding unit 104.
- the object position attribute information stored in the object position attribute holding unit 104 is information shown in FIG.
- the gripping target information acquisition unit 105 acquires the position of the gripping target object, and outputs the position to the secondary collision possibility direction evaluation unit 108, the secondary collision possibility distance evaluation unit 109, and the route evaluation unit 113. Further, the gripping target information acquisition unit 105 outputs the attribute information of the gripping target object to the risk attribute combination evaluation unit 112 (step S703).
- the gripping target information acquisition unit 105 sets the position (40, 35, 0) to the secondary collision possibility direction. The result is output to the evaluation unit 108, the secondary collision possibility distance evaluation unit 109, and the route evaluation unit 113. Further, the gripping target information acquisition unit 105 outputs the attribute information “garlic” of the gripping target object to the risk attribute combination evaluation unit 112.
- the path evaluation unit 113 determines whether there is an object that has not been acquired as a secondary collision candidate object among the objects corresponding to the attribute information held by the object position attribute holding unit 104 (step S704).
- the path evaluation unit 113 arbitrarily selects one of the unacquired objects, and selects the selected object as a secondary object. Let it be a collision candidate object.
- the robot and the gripping target object itself are not secondary collision candidate objects. Specifically, in the case of FIG. 3, for example, “magnet” and “cake” are the secondary collision candidate objects.
- the secondary collision possibility direction evaluation unit 108 calculates the secondary collision possibility direction evaluation value of the object designated as the secondary collision candidate object, and outputs it to the route evaluation unit 113 (step S706).
- the vector connecting the positions of “garlic” and “cake” is (50-40, 15-35, 0), that is, the vector (10, -20, 0).
- the movement path connecting the position (40, 35, 0) of the “garlic” that is the object to be grasped and the position (10, 35, 0) of the robot is a vector (30, 0, 0). Therefore, the inner product of these vectors is calculated by the following equation 1.
- Equation 4 Since the inner product of the two vectors is equal to the product of the length of each vector multiplied by the cosine of the angle formed by the vector, Equation 4 is established.
- the secondary collision possibility distance evaluation unit 109 calculates the secondary collision possibility distance evaluation value of the object designated as the secondary collision candidate object, and outputs it to the route evaluation unit 113 (step S707).
- the secondary collision possibility distance evaluation value is e.
- the danger attribute combination evaluation unit 112 calculates the danger attribute combination evaluation value of the secondary collision candidate object and outputs it to the route evaluation unit 113 (step S709).
- the route evaluation unit 113 calculates the secondary collision possibility direction evaluation value calculated in the secondary collision possibility direction evaluation process (step S706) and the secondary collision calculated in the secondary collision possibility distance evaluation process (step S707). Based on the collision possibility distance evaluation value and the risk attribute combination evaluation value calculated in the risk attribute combination acquisition process (step S709), the evaluation value of the secondary collision candidate object is calculated. That is, the damage risk when a secondary collision occurs is calculated (step S710).
- step S710 After the evaluation calculation process (step S710), the process returns to the secondary collision candidate object determination process (step S704), and the same process is repeated for the remaining secondary collision candidate objects.
- the path evaluation unit 113 determines a path evaluation value based on the evaluation of each candidate calculated in the evaluation calculation process (step S710). Is calculated. The calculated route evaluation value is stored in the evaluation holding unit 114 (step S711).
- the route evaluation unit 113 evaluates the evaluation value when “cake” is a secondary collision candidate object.
- the route evaluation value is obtained by calculating the sum with the evaluation value 0 when the “magnet” calculated in the same manner is a secondary collision candidate object.
- the route evaluation value is
- the action planning unit 115 acquires a route evaluation value from the evaluation holding unit 114, executes an action plan based on the acquired route evaluation value, and stores information indicating the planned action in the plan holding unit 116 (step). S712).
- information indicating a low speed is stored as the moving speed of the robot 401, and information indicating a high speed is stored when the value is equal to or lower than the predetermined value.
- the robot position acquisition unit 101 has acquired the robot position using GPS or image processing, but other methods may be used.
- the position of the robot may be determined based on the initial position of the robot and the distance and direction in which the robot is moved from the initial position to the time when the action plan is started.
- the attribute information is information indicating the type of the object.
- the attribute information may be information related to properties such as “strong smell” and “emits magnetic force”.
- the dangerous attribute combination information holding unit 111 holds a dangerous attribute combination of “strong smell, cake”.
- the attribute information of “garlic”, “ginger”, and “wasabi” is “strong smell”, it can be expressed by one risk attribute combination that they are dangerous for “cake”.
- the attribute information in the dangerous attribute combination may be a concept higher than the attribute information given to the object.
- the dangerous attribute combination information holding unit 111 holds a dangerous attribute combination of “food with strong smell”.
- the upper attribute information of “cake” and “strawberry” is “food”, it can be expressed by one risk attribute combination that they are dangerous for “scented” items.
- the attribute information is information indicating the type of the object.
- information such as “weak to“ magnetic ”” and “weak to“ strong smell ”” may be used.
- ⁇ Modification 4> the size of the object to be grasped is not considered. Therefore, when grasping an object, a route that moves to the position of the object to be grasped is planned. However, in practice, since the object has a size, it is possible to make a plan to stop the movement slightly before the object to be grasped in consideration of the size.
- the route evaluation unit 113 evaluates each route, and the evaluation holding unit 114 holds the evaluation for each evaluation.
- the action plan unit 115 performs an action plan for each route.
- the plan holding unit 116 holds each action plan.
- a well-known technique is used for the production
- the path evaluation unit 113 calculates an object risk in the environment as a secondary collision candidate, calculates a damage risk for each candidate, and uses the total value as a path evaluation value.
- the maximum damage risk may be used as the route evaluation value.
- the secondary collision possibility direction evaluation value, the secondary collision possibility distance evaluation value, and the risk attribute combination evaluation value are calculated in order, but these values are equal to or less than a predetermined value. In some cases, the calculation of evaluation values after that may be omitted. If such a structure is used, calculation time can be shortened.
- the risk attribute combination evaluation value may be calculated first, and when there are few objects, the risk attribute combination evaluation value may be calculated last. If such a configuration is used, the calculation time can be shortened.
- a route including a plurality of directed line segments may be generated and evaluation and action planning may be performed for each route as in the fifth modification.
- the secondary collision possibility distance evaluation unit 109 calculates the distance between the object to be grasped and the surrounding object
- a configuration may be adopted in which a plurality of routes are formed by the above method and the total distance of the plurality of routes is used. Absent.
- the secondary collision possibility direction evaluation unit 108 and the secondary collision possibility distance evaluation unit 109 perform the secondary collision possibility direction evaluation value and the secondary collision possibility without considering the size of the object.
- the distance evaluation value is calculated.
- the configuration may be such that the secondary collision possibility direction evaluation value and the secondary collision possibility distance evaluation value are calculated using an equation that takes into account the size of the object. This is because, when two objects exist at the same distance from the object to be grasped, the larger the size of the object, the higher the possibility of causing a secondary collision.
- the size of the object may be stored in the object position attribute holding unit 104.
- the circumferential length of the gripping target object 1801 having the radius r is 2 ⁇ ⁇ ⁇ r.
- the peripheral object 1802 occupies a length of about 2 ⁇ s. Therefore, when the friction coefficient of the plane is ignored, a secondary collision occurs with a probability of s / ( ⁇ ⁇ r). Therefore, s / ( ⁇ ⁇ r) may be used as the secondary collision possibility distance evaluation value.
- the radius S1 of the circumscribed circle of the surrounding object is used instead of the radius s.
- the radius of the circumscribed circle may be acquired by image processing, or may be described in advance on a tag attached to the object and read by a tag reader when planning an action.
- the secondary collision possibility direction evaluation unit 108 and the secondary collision possibility distance evaluation unit 109 perform the secondary collision possibility direction evaluation value and the secondary collision possibility without considering the size of the surrounding object.
- the sex distance evaluation value was calculated.
- the secondary collision possibility direction evaluation unit 108 and the secondary collision possibility distance evaluation unit 109 use a formula that takes into account the size of the surrounding object, and uses the secondary collision possibility direction evaluation value and the secondary collision possibility distance.
- the configuration may be such that the evaluation value is calculated. This is because, when two objects exist at the same distance from the object to be grasped, the larger the size of the object, the higher the possibility of causing a secondary collision.
- the size of the object may be stored in the object position attribute holding unit 104.
- FIG. 9 there will be described a case where there is a peripheral object 1802 with a radius s at a distance of radius r from the object to be gripped 1801 and the movement path 1803 of the robot is determined as shown in FIG.
- the peripheral object 1802 exists on the straight line 1804 passing through the center position of the gripping target object 1801 obtained by translating the movement path 1803, the secondary of the gripping target object 1801 and the peripheral object 1802 is obtained. Collisions are likely to occur.
- an angle between a straight line connecting the centers of the gripping target object 1801 and the peripheral object 1802 and the straight line 1804 is ⁇ .
- an angle formed by a straight line connecting the centers of the gripping target object 1801 and the peripheral object 1802 and a tangent of a circumscribed circle of the peripheral object 1802 passing through the gripping target object 1801 is ⁇ . That is, the tangent of the circumscribed circle of the peripheral object 1802 passing through the gripping target object 1801 indicates a boundary line whether or not the gripping target object 1801 collides with the peripheral object 1802, and ⁇ indicates a boundary angle.
- ⁇ is calculated according to the following formula.
- the case where the secondary collision is likely to occur can be considered as a case where the relationship of ⁇ ⁇
- the secondary collision possibility direction evaluation unit 108 and the secondary collision possibility distance evaluation unit 109 perform the secondary collision possibility direction without considering the size of the object to be grasped and the size of the surrounding objects.
- the evaluation value and the secondary collision possibility distance evaluation value are calculated.
- the secondary collision possibility direction evaluation unit 108 and the secondary collision possibility distance evaluation unit 109 use a formula that considers the size of both objects, and evaluate the secondary collision possibility direction evaluation value and the secondary collision possibility distance.
- the configuration may be such that the evaluation value is calculated. This is because when two objects exist at the same distance from the gripping target object, the larger the size of the object, the more likely it is to cause a secondary collision, and the larger the size of the gripping target object, the higher the secondary collision. It is because there is a high possibility of causing.
- the size of the object may be stored in the object position attribute holding unit 104.
- the radius of the circumscribed circle of the gripping target object 1901 is t
- the radius of the circumscribed circle of the peripheral object 1802 is s
- the distance between the center of the gripping target object 1901 and the center of the peripheral object 1802 is r.
- the movement path 1803 of the robot is determined as shown in FIG.
- a tangent common to the gripping target object 1901 and the peripheral object 1802 indicates a boundary line indicating whether or not the gripping target object 1901 collides with the peripheral object 1802.
- an angle ⁇ formed by the tangent and a straight line connecting the centers of the gripping target object 1901 and the peripheral object 1802 indicates a boundary angle. Note that ⁇ is calculated according to the following equation.
- an angle between a straight line connecting the centers of the gripping target object 1901 and the peripheral object 1802 and a straight line 1804 passing through the center position of the gripping target object 1901 obtained by translating the movement path 1803 is defined as ⁇ .
- the secondary collision possibility direction evaluation value is set to 1.
- both the secondary collision possibility direction evaluation value and the secondary collision possibility distance evaluation value are set to 1.
- the secondary object in addition to the primary damage caused by the object that the mobile robot touches or approaches, the secondary object is detected when the object further contacts the surrounding object. It can also be determined whether or not damage will occur. Thereby, the action plan of a mobile robot can be performed. Along with this, secondary damage can be reduced.
- the robot action plan for avoiding the damage risk is performed.
- the robot can be moved at high speed, so that the task execution time can be reduced. Further, by moving the robot at high speed, the time for supplying electric power to the robot is reduced, so that the energy cost can be reduced.
- the robot can be moved at high speed while avoiding constant movement of the robot at a low speed. There is no. When a person observing the movement of the robot feels annoyed, the person himself moves the object. By doing so, the energy cost used until the robot is carrying the object is wasted, but the irritability can be avoided and the effect is great.
- FIG. 11 is a diagram showing the robot system 805.
- a robot system 805 is a robot system that supports a user's cooking work.
- the robot system 805 includes a work table 803, a robot 800 that moves on the work table 803, and an action change instruction device 802 that is used when a user instructs the robot 800 to perform an emergency stop or change work.
- the robot 800 is a robot that moves on the work table 803, grips an object instructed by the user, and carries the gripped object to the place where the user is.
- the robot 800 includes an action planning device 100 and a moving device 801.
- the behavior planning apparatus 100 plans the behavior described in the first embodiment. That is, the moving action on the work table 803 for holding the object instructed by the user is planned.
- the action planning apparatus 100 is provided inside the robot 800.
- the moving device 801 is a device that moves the robot 800 in accordance with the action plan planned by the action planning apparatus 100.
- the object position attribute acquisition unit 103 that is a part of the behavior planning apparatus 100 is a processing unit that acquires object position attribute information on the work table 803. Some of them are fixed to the environment and are not moved by the moving device 801.
- the action planning apparatus 100 makes an action plan similar to that of the first embodiment, and the robot 800 performs an action according to the action plan.
- the user can operate the behavior change instruction device 802 to urgently stop the robot 800 or change the object to be grasped.
- the robot can be moved at a low speed when there is a risk of damage from a secondary collision. Therefore, the time required for the robot to reach the object to be grasped is longer than when moving at high speed. For this reason, even when the robot is moving along a dangerous route that is likely to cause a secondary collision, the user can be aware of this, and the action change instruction device 802 can stop the robot.
- FIG. 12 is a diagram showing the robot arm 901.
- the robot arm 901 moves the robot arm by controlling the joint of the robot arm 901 based on the behavior planned by the behavior planning apparatus 100a.
- each joint of the arm moves to a different position.
- the behavior planning apparatus 100a uses each joint as a representative point of the robot arm 901, calculates a damage risk for each representative point, and uses the total as the total amount of damage risk.
- FIG. 13 is a block diagram showing the configuration of the action planning apparatus 100a according to the third embodiment.
- the difference between the action planning apparatus 100a and the action planning apparatus 100 according to the first embodiment will be described.
- a position holding unit 102a, a route generation unit 106a, a route holding unit 107a, and a route evaluation unit 113a are used.
- Other configurations are the same as those of the action planning apparatus 100.
- a part of the object position attribute acquisition unit 103 is fixed to the environment of the robot arm 901.
- the robot position acquisition unit 101a is a processing unit that acquires the position of each representative point of the robot arm and stores the acquired position of the representative point in the start point position holding unit 102a.
- the acquisition method is the same as the robot position acquisition method by the robot position acquisition unit 101 of the first embodiment.
- the representative points of the robot arm 901 are the tip of the robot arm 901 and the position of each joint.
- the start point position holding unit 102a is a storage device that stores the position of each representative point of the robot arm 901 acquired by the robot position acquisition unit 101a, and is configured by, for example, an HDD.
- the path generation unit 106a is a processing unit that generates a path for moving the representative point at the tip of the robot arm 901 to the position of the object to be grasped. Specifically, for example, a directed straight line having a representative point at the tip of the robot arm 901 as a start point and an end point at a position of the gripping target object is set as a route. Further, the path generation unit 106 a generates a path of representative points other than the representative point at the tip of the robot arm 901. A known method related to the movement of the articulated robot arm is used to generate the path.
- the route holding unit 107a is a storage device that holds a set of routes of representative points generated by the route generation unit 106a, and is configured by, for example, an HDD.
- the route evaluation unit 113a calculates a damage risk when the robot arm 901 moves according to a set of routes held by the route holding unit 107a.
- the route evaluation unit 113 calculates one evaluation value for one route, whereas the route evaluation unit 113a differs in that one evaluation value is calculated for a set of routes. That is, the route evaluation unit 113a calculates a route evaluation value (damage risk) for each representative point route by the same method as the route evaluation unit 113, and uses the sum as an evaluation value for the set of routes.
- the route evaluation unit 113a stores the calculated route evaluation value in the evaluation holding unit 114.
- tip of the arm and each joint are used as representative points here, other positions may be used as representative points.
- the action planning apparatus 100a makes an action plan similar to that of the first embodiment, and the robot arm 901 performs an action according to the action plan.
- the behavior planning apparatus has been described in which a behavior plan for a robot is drafted on the assumption that a secondary collision occurs based on a primary collision between the robot and the object to be grasped.
- the robot may collide first with an object on the moving path of the robot other than the object to be grasped. The reason is that, as with the collision with the object to be grasped, the position recognition accuracy of the robot in the home environment is low, the position control accuracy of the robot is low, or the task execution time is severely limited. Because.
- an action planning apparatus 100b that makes a robot action plan in consideration of such a primary collision with an object other than the object to be grasped will be described.
- the action plan in the case where the object to be grasped and the robot perform a primary collision and a secondary collision occurs has been described.
- the robot moves, there is a possibility that an object existing on the movement path and the robot collide first.
- the action planning apparatus 100b according to the fourth embodiment uses an object in the environment as a candidate for a primary collision with the robot, and the closer the object from the movement path of the robot (and the robot recognizes) the closer the robot is to the robot. It is evaluated that a primary collision is likely to occur.
- an object with a longer distance is less likely to cause a primary collision with the robot.
- another object in the environment is further selected as a candidate for a secondary collision, and whether or not the secondary collision is likely to occur and the degree of damage when it occurs are evaluated.
- the action planning apparatus 100b determines the risk of damage that causes a secondary collision between an object that is a candidate for a primary collision and an object that is a candidate for a secondary collision. And the product of the degree of damage when a secondary collision occurs. Thereby, when the calculated damage risk is larger than a predetermined value, an action for avoiding the secondary collision is planned.
- FIG. 14 is a block diagram illustrating a configuration of the action planning apparatus 100b according to the fourth embodiment.
- the behavior planning apparatus 100b instead of the route generation unit 106, the secondary collision possibility direction evaluation unit 108, the secondary collision possibility distance evaluation unit 109, the risk attribute combination evaluation unit 112, and the route evaluation unit 113 in the action planning device 100.
- the route generation unit 106b, the secondary collision possibility direction evaluation unit 108b, the secondary collision possibility distance evaluation unit 109b, the risk attribute combination evaluation unit 112b, and the route evaluation unit 113b are used.
- the action planning device 100b includes a primary collision possibility evaluation unit 1101.
- the route generation unit 106b is a processing unit that acquires a movement destination of the robot from the outside and generates a route.
- the path generation unit 106b includes an input device (not shown), for example, and acquires the three-dimensional coordinates input by the user as the movement destination.
- the path generation unit 106b uses a directed line segment starting from the robot position that is the starting point of the moving path of the robot held by the start point position holding unit 102 and having the acquired moving destination as the end point.
- the secondary collision possibility direction evaluation unit 108b is a processing unit that evaluates how easily the designated secondary collision candidate object is in a collision direction with respect to the designated primary collision candidate object.
- the secondary collision possibility direction evaluation unit 108b is different in that a primary collision candidate object is used instead of the gripping target object in the secondary collision possibility direction evaluation unit 108.
- the secondary collision possibility distance evaluation unit 109b is a processing unit that evaluates how easily the designated secondary collision candidate object is at a collision distance with respect to the designated primary collision candidate object.
- the secondary collision possibility distance evaluation unit 109b is different in that a primary collision candidate object is used instead of the gripping target object in the secondary collision possibility distance evaluation unit 109.
- the dangerous attribute combination evaluation unit 112b is a processing unit that evaluates whether the designated secondary collision candidate object is a dangerous object with respect to the designated primary collision candidate object.
- the danger attribute combination evaluation unit 112b is different in that a primary collision candidate object is used instead of the gripping target object in the danger attribute combination evaluation unit 112.
- the primary collision possibility evaluation unit 1101 is a processing unit that evaluates how easily the designated primary collision candidate object is located on the movement path of the robot. In the following description, this evaluation is called “primary collision possibility evaluation”, and the evaluation value calculated at that time is called “primary collision possibility evaluation value”.
- This primary collision possibility evaluation is an evaluation of the ease of occurrence of the primary collision described in the above ⁇ Concept of avoiding damage due to secondary collision in moving behavior>.
- the primary collision possibility evaluation unit 1101 sets the distance between the robot movement path and the specified object to r3, and uses the Euler number e to the ⁇ r3 power (e ⁇ r3 ) as the primary collision possibility evaluation value.
- the primary collision possibility evaluation value when the distance is 0, the primary collision possibility evaluation value is the maximum value 1, and when the distance is infinite, the primary collision possibility evaluation value is the minimum value 0. That is, the primary collision possibility evaluation value is closer to the maximum value of 1 as the specified object is more likely to collide with the route, and is the minimum value as the distance is less likely to collide. The value is close to zero.
- the expression for evaluating the possibility of primary collision here may be another expression as long as the value decreases as the distance increases. Also, an equation according to a model that takes into account the appearance characteristics of sensor errors may be used.
- the route evaluation unit 113b is a processing unit that evaluates the damage risk due to the secondary collision when the robot moves on the route held by the route holding unit 107.
- the route evaluation unit 113b evaluates the damage risk according to the concept described in the above ⁇ Concept of avoiding damage due to secondary collision in moving behavior>.
- the robot position acquisition unit 101 acquires the robot position and stores it in the start point position holding unit 102 (step S701).
- the object position attribute acquisition unit 103 acquires object position attribute information of an object existing in the environment, and stores it in the object position attribute holding unit 104 (step S702).
- the robot position acquisition process (step S701) and the object position attribute information acquisition process (step S702) are the same as those described in the first embodiment.
- the route evaluation unit 113b sets the value of the evaluation storage variable to 0 (step S1200).
- the path evaluation unit 113b determines whether there is an object that has not been acquired as a primary collision candidate object among the objects corresponding to the attribute information held by the object position attribute holding unit 104 (step S1201). If there is an unacquired object as the primary collision candidate object (YES in step S1201), the path evaluation unit 113b arbitrarily selects one of the unacquired objects.
- the primary collision possibility evaluation unit 1101 calculates a primary collision possibility evaluation value for the selected primary collision candidate object (step S1202).
- the path evaluation unit 113b determines whether there is an object that is not acquired as a secondary collision candidate object among objects other than the primary collision candidate object corresponding to the attribute information held by the object position attribute holding unit 104. (Step S1203). If there is an unacquired object as a secondary collision candidate object (YES in step S1203), the path evaluation unit 113b arbitrarily selects one of the unacquired objects, and selects the selected object as a secondary object. Let it be a collision candidate object.
- the secondary collision possibility direction evaluation unit 108b calculates a secondary collision possibility direction evaluation value indicating a possibility that the primary collision candidate object and the secondary collision candidate object collide secondarily (step S1204).
- the secondary collision possibility distance evaluation unit 109b calculates a secondary collision possibility distance evaluation value indicating the possibility that the primary collision candidate object and the secondary collision candidate object will collide with each other (step S1205).
- the risk attribute combination evaluation unit 112b calculates a risk attribute combination evaluation value indicating the possibility of damage when the primary collision candidate object and the secondary collision candidate object collide with each other (step S1206).
- the secondary collision possibility direction evaluation process (step S1204), the secondary collision possibility distance evaluation process (step S1205), and the danger attribute combination acquisition process (step S1206) are the secondary collision possibility direction shown in the first embodiment.
- the evaluation process (step S706) is the same as the secondary collision possibility distance evaluation process (step S707) and the danger attribute combination acquisition process (step S709). However, the difference is that a primary collision candidate object is used instead of the gripping target object.
- the path evaluation unit 113b calculates the damage risk for the combination of the primary collision candidate object and the secondary collision candidate object, and stores the addition result of the calculation result and the value of the evaluation storage variable in the evaluation storage variable ( Step S1207).
- the damage risk is (secondary collision possibility direction evaluation value) ⁇ (secondary collision possibility distance evaluation value) ⁇ (primary collision possibility evaluation value) ⁇ (danger attribute combination evaluation value).
- step S1207 After the evaluation calculation process (step S1207), the process returns to the secondary collision candidate object determination process (step S1203), and the same process is repeated for the remaining secondary collision candidate objects.
- step S1203 If there are no secondary collision candidate objects (NO in step S1203), the process returns to the primary collision candidate object determination process (step S1201), and the same process is repeated for the remaining primary collision candidate objects. It is. Every time the primary collision candidate object changes, all other objects are returned to the unacquired state as secondary collision candidate objects.
- the path evaluation unit 113b evaluates the value stored in the evaluation storage variable as a path evaluation.
- the value is stored in the evaluation holding unit 114 (step S1208).
- the action planning unit 115 acquires a route evaluation value from the evaluation holding unit 114, executes an action plan based on the acquired route evaluation value, and plans information indicating the planned action.
- the data is stored in the holding unit 116 (step S712).
- ⁇ Modification 1> the risk evaluation method for the path by calculating the damage risk that the first object on the moving path of the robot and the second object around the robot collide secondarily has been described. Furthermore, in the fourth embodiment, the action plan for planning the moving speed of the robot near the first object based on the damage risk has been described.
- the processing performed based on the damage risk evaluation is not limited to this, and any processing may be performed.
- the behavior planning apparatus 100c plans a route on which the robot moves to the destination as a collection of a plurality of routes.
- the movement path of the robot 401 is divided into three paths (paths 1301, 1302). And 1303) is an example of a case where the route group is planned.
- the behavior planning apparatus 100c calculates a damage risk for each designed route, and when there is a route with a high damage risk, plans a new route that does not pass through the route.
- the action planning device 100c plans a new alternative route as shown in FIG. That is, for example, the behavior planning apparatus 100c assumes that there is an obstacle 1308 at the position of the route 1302 in FIG. 16, and generates another route by using a conventional route generation method in an environment with an obstacle. That is, a route is generated when there is an obstacle 1308 such as a wall, and the generated route is set as another route.
- routes 1301, 1305, 1306, 1307, and 1303 are generated as separate routes.
- the behavior planning apparatus 100c also evaluates whether there is a route with a high risk of damage from the secondary collision even with the newly generated another route, and if there is, further generates another route excluding that portion.
- FIG. 18 is a block diagram illustrating a configuration of an action planning apparatus 100c according to the fifth embodiment.
- the behavior planning apparatus 100c uses an excluded area holding unit 1501.
- the excluded area holding unit 1501 is a storage device that holds an area to be excluded when a route is generated, and includes an HDD or the like. In the initial state, nothing is stored in the excluded area holding unit 1501.
- the route generation unit 106c is a processing unit that acquires a robot destination from the outside and generates a route.
- the route generation unit 106c includes, for example, an input device (not shown), and acquires the three-dimensional coordinates input by the user as the movement destination.
- the route generation unit 106b according to the fourth embodiment generates one route
- the route generation unit 106c generates a route group including a plurality of directed line segments.
- the route generation is performed assuming that there is an obstacle in the region held by the excluded region holding unit 1501.
- the route holding unit 107c is a storage device that holds a plurality of routes generated by the route generation unit 106c, and includes, for example, an HDD.
- the evaluation holding unit 114c is a storage device that holds, as a route evaluation value, an evaluation calculated as a damage risk for each route by the route evaluation unit 113, and is configured by, for example, an HDD. Whereas the evaluation holding unit 114 in the fourth embodiment holds one evaluation, the evaluation holding unit 114c holds an evaluation for each route held by the route holding unit 107c.
- the behavior planning unit 115c determines whether there is a route evaluation value equal to or greater than a predetermined value among the route evaluation values held in the evaluation holding unit 114c. If there is a route evaluation value greater than or equal to the predetermined value, the action planning unit 115c acquires a route corresponding to the route evaluation value from the route holding unit 107c. In addition, the action planning unit 115c adds an area of a circle (or a sphere or a cylinder) whose diameter is the start point and the end point of the acquired route to the excluded area holding unit 1501 as an excluded area. When the excluded area is added, a new route is generated, and the route evaluation unit 113 performs route evaluation again. If there is no route evaluation value equal to or greater than the predetermined value, the behavior planning unit 115c outputs the route group held by the route holding unit 107c to the plan holding unit 116c.
- the plan holding unit 116c is a storage device that holds information indicating the action plan planned by the action planning unit 115c, and includes, for example, an HDD.
- the action planning apparatus 100c makes an action plan similar to that of the fourth embodiment while changing the movement path of the robot, and the robot performs an action according to the action plan.
- the route generation unit 106c is designed as a plurality of routes from the beginning to the destination of the robot, but may be a plurality of routes as long as it can be a single straight route. It doesn't matter.
- the damage risk is calculated as a straight line, and if the damage risk is lower than a predetermined value, the straight line is taken as a route. If the damage risk is higher than a predetermined value, the route is subdivided to determine which part of the route It is also possible to find out whether the risk of damage is high and design an improvement route by excluding that part. If such a configuration is used, the calculation load for calculating the damage risk can be reduced.
- the action plan unit 115c uses a circle (or sphere) region having a diameter at the start point and the end point of the route as an excluded region, but may have other shapes. Absent. In other words, if there are many routes with a high risk of damage, it may be possible to prevent the improvement route from being recreated many times by taking a larger area to be excluded.
- the improved route can be made to make a route that does not go around so much.
- the area of the area to be excluded may be determined based on the path evaluation value. For example, when the path evaluation value is large (that is, when the damage risk is high), the area of the excluded area is increased, and when the path evaluation value is small (that is, when the damage risk is low), the area of the excluded area It is possible to make the configuration smaller. By using such a configuration, when the risk of damage is small, it is possible to find a travel route as short as possible compared to always excluding a large area. Conversely, when the damage risk is high, a route with a low damage risk can be found with a smaller number of times compared to always excluding a small area.
- ⁇ Modification 3> another route is planned for the action planning device 100b shown in the fourth embodiment. However, another route is planned for the behavior planning device 100 shown in the first embodiment. You may make it do. That is, the primary collision possibility evaluation unit 1101 is not provided, and the secondary collision possibility direction evaluation unit 108b, the secondary collision possibility distance evaluation unit 109b, and the risk attribute combination evaluation unit 112b are replaced with the secondary collision. The possibility direction evaluation unit 108, the secondary collision possibility distance evaluation unit 109, and the risk attribute combination evaluation unit 112 may be used.
- the damage risk can be reduced by changing the movement path of the robot, and the effect is great.
- the behavior planning apparatus 100b that plans the behavior of moving at a low speed when the risk of damage from the secondary collision is high is shown. However, when the robot executes a task, a plurality of movement paths may be assumed. In the sixth embodiment, an action planning apparatus 100d that selects an action with a low risk of damage from a secondary collision among a plurality of action plans will be described.
- FIG. 19 is a diagram illustrating an example when a plurality of routes are assumed.
- the robot holds garlic
- FIG. 20 is a block diagram illustrating a configuration of the action planning apparatus 100d according to the sixth embodiment.
- the difference between the action planning apparatus 100d and the action planning apparatus 100b according to the fourth embodiment will be described.
- the action planning device 100d instead of the route generation unit 106b, the route holding unit 107, the evaluation holding unit 114, the behavior planning unit 115, and the plan holding unit 116, the route generation unit 106d, the route holding unit 107c, and the evaluation holding unit 114c, respectively.
- the action planning unit 115d and the plan holding unit 116d are used.
- the route generation unit 106d plans the movement behavior of the robot to the position of the designated object. For example, when “garlic” is designated, the path generation unit 106 d acquires the position of “garlic” in the environment from the object position attribute holding unit 104. When a plurality of designated objects are registered in the object position attribute holding unit 104, the route generation unit 106d plans a route for each object and holds the route in the route holding unit 107c.
- the route holding unit 107c holds a plurality of routes generated by the route generation unit 106d as in the fifth embodiment.
- the evaluation holding unit 114c holds the evaluation calculated by the route evaluation unit 113 as the damage risk for each route, as in the fifth embodiment.
- the behavior planning unit 115d selects the evaluation with the lowest evaluation, that is, the one with the lowest damage risk among the evaluations held by the evaluation holding unit 114c, and holds the selected route in the plan holding unit 116d.
- the plan holding unit 116d is a storage device that holds the route generated by the behavior planning unit 115d, that is, the robot's behavior plan, and includes, for example, an HDD.
- ⁇ Modification 1> when there are a plurality of objects having the same name and there are a plurality of movement paths, a case where a path with a low damage risk is selected is described.
- the present invention is not limited to this.
- the path with the lowest damage risk may be selected from the paths to the respective gripping points.
- the route generation unit 106d generates a plurality of routes in advance.
- a configuration for planning a route may be used.
- the action planning unit 115d in the sixth embodiment is configured to select a route with the least damage risk, but may be configured to perform route selection in consideration of another factor. For example, since a route having a long movement route length consumes extra costs such as energy and time, the route evaluation may be performed by combining the damage risk and the route length.
- the risk of the moving path for the robot is evaluated based on the damage risk due to the secondary collision.
- a delicate glass vase may be damaged by a primary collision with the robot, and there is a risk of damage due to the damage.
- Embodiment 7 describes an action planning apparatus that evaluates the danger of a moving route in consideration of damage risk caused by a primary collision and damage risk caused by a secondary collision, and performs a robot action plan.
- the damage risk due to the primary collision is evaluated based on the positional relationship between the robot and the object in the environment and the attribute information of the object in the environment. That is, the damage risk due to the primary collision is evaluated from the evaluation result of the primary collision possibility evaluation and the evaluation result of the primary collision risk attribute evaluation.
- the primary collision possibility evaluation is an evaluation of whether an object in the environment is present at a position where a collision is likely to occur with respect to the movement path of the robot.
- the primary collision risk attribute evaluation is an evaluation of whether or not an object is likely to generate damage risk when a collision between the robot and an object in the environment occurs.
- FIG. 21 is a block diagram illustrating a configuration of an action planning apparatus 100e according to the seventh embodiment.
- the action planning device 100e uses the route evaluation unit 113e instead of the route evaluation unit 113b in the action planning device 100b.
- the behavior planning apparatus 100e includes a primary collision risk attribute information holding unit 2101 and a primary collision risk attribute evaluation unit 2102 that are not included in the behavior planning apparatus 100b.
- the primary collision risk attribute information holding unit 2101 is primary collision risk attribute information that is information indicating that damage may occur when a robot collides with an object having what kind of attribute information. Is remembered.
- FIG. 22 is a diagram illustrating an example of primary collision risk attribute information. Objects such as glasswork and plastic models are fragile due to contact with the robot (even if secondary collisions do not occur). In addition, objects such as flower pots, drink cups, and flour may be diffused by contact with the robot.
- the primary collision risk attribute evaluation unit 2102 is a processing unit that evaluates whether a specified object is an object that may cause damage in the case of a primary collision with a robot. In the following description, this evaluation is called “primary collision risk attribute evaluation”, and the evaluation value calculated at that time is called “primary collision risk attribute evaluation value”.
- the primary collision risk attribute evaluation is an evaluation for estimating the degree of damage when a primary collision occurs. Even if a robot collides with an object in the environment, damage does not always occur. For example, even if the robot collides with the clip, the damage hardly occurs. Even if damage occurs, the degree of damage is different. For example, if a robot collides with a glasswork, and the glasswork breaks, the price of the glasswork and the time and man-hour to clean the shattered glasswork will be the amount of damage, which is sometimes enormous. Damage. On the other hand, as a case where the damage is small, a case where the robot and the spoon collide can be considered.
- the degree of damage is small. As described above, although the degree of damage varies, in the seventh embodiment, the degree of damage is evaluated in two stages. That is, the primary collision risk attribute evaluation unit 2102 classifies the damage degree into two cases: a case where the damage degree is small enough to be ignored and a case where the damage degree is too large to be ignored.
- the action planning unit 115 performs an action plan for a robot that does not perform primary collision avoidance when the degree of damage is small.
- the primary collision risk attribute evaluation unit 2102 is in the case of a primary collision with the designated object.
- the degree of damage is large. Further, when the attribute information of the designated object is not stored in the primary collision risk attribute information holding unit 2101, the primary collision risk attribute evaluation unit 2102 is in the time of the primary collision with the designated object. It is judged that the degree of damage is small.
- the primary collision risk attribute evaluation unit 2102 acquires the attribute information of the designated object from the object position attribute holding unit 104. The primary collision risk attribute evaluation unit 2102 determines whether the acquired attribute information is included in the attribute information held by the primary collision risk attribute information holding unit 2101. The primary collision risk attribute evaluation unit 2102 outputs the determination result as a primary collision risk attribute evaluation value to the route evaluation unit 113e.
- the determination result is “1” when included in the primary collision risk attribute information, and is “0” when not included.
- the primary collision risk attribute information holding unit 2101 stores, for each of the primary collision risk attribute information shown in FIG.
- the damage amount when the object having the corresponding attribute information collides with the robot. good if the attribute information of the designated object is included in the attribute information held by the primary collision risk attribute information holding unit 2101, the damage amount of the attribute information is set as the primary collision risk attribute evaluation value. If not included, “0” is set as the primary collision risk attribute evaluation value.
- the route evaluation unit 113e is a processing unit that evaluates the damage risk when the robot moves on the route held by the route holding unit 107.
- the route evaluation unit 113e evaluates the route damage risk based on the damage risk of the primary collision and the secondary collision.
- FIG. 23 shows a flow of action plan processing performed by the action plan apparatus 100e.
- the robot position acquisition unit 101 acquires the robot position and stores it in the start point position holding unit 102 (step S701).
- the object position attribute acquisition unit 103 acquires object position attribute information of an object existing in the environment, and stores it in the object position attribute holding unit 104 (step S702).
- the robot position acquisition process (step S701) and the object position attribute information acquisition process (step S702) are the same as those described in the first embodiment.
- the route evaluation unit 113e sets the value of the evaluation storage variable to 0 (step S1200).
- the route evaluation unit 113e determines whether there is an object that has not been acquired as a primary collision candidate object among the objects corresponding to the attribute information held by the object position attribute holding unit 104 (step S1201). If there is an unacquired object as the primary collision candidate object (YES in step S1201), the path evaluation unit 113e arbitrarily selects one of the unacquired objects.
- the primary collision possibility evaluation unit 1101 calculates a primary collision possibility evaluation value for the selected primary collision candidate object (step S1202).
- the primary collision risk attribute evaluation unit 2102 calculates a primary collision risk attribute evaluation value of the selected primary collision candidate object (step S2302).
- the route evaluation unit 113e calculates a primary collision route evaluation value.
- the primary collision path evaluation value is a value indicating a damage risk when the moving object and the selected object collide first, and here, the primary collision possibility evaluation value and the primary collision risk attribute.
- the product with the evaluation value is a value indicating a damage risk when the moving object and the selected object collide first.
- the route evaluation unit 113e adds the calculated primary collision route evaluation value and the value of the evaluation storage variable, and stores the addition result in the evaluation storage variable. (Step S2303).
- the path evaluation unit 113e determines whether there is an object that is not acquired as a secondary collision candidate object among objects other than the primary collision candidate object corresponding to the attribute information held by the object position attribute holding unit 104. (Step S1203). If there is an unacquired object as a secondary collision candidate object (YES in step S1203), the path evaluation unit 113e arbitrarily selects one of the unacquired objects, and selects the selected object as a secondary object. Let it be a collision candidate object.
- the secondary collision possibility direction evaluation unit 108b calculates a secondary collision possibility direction evaluation value indicating a possibility that the primary collision candidate object and the secondary collision candidate object collide secondarily (step S1204).
- the secondary collision possibility distance evaluation unit 109b calculates a secondary collision possibility distance evaluation value indicating the possibility that the primary collision candidate object and the secondary collision candidate object will collide with each other (step S1205).
- the risk attribute combination evaluation unit 112b calculates a risk attribute combination evaluation value indicating the possibility of damage when the primary collision candidate object and the secondary collision candidate object collide with each other (step S1206).
- the path evaluation unit 113e calculates a secondary collision path evaluation value for the combination of the primary collision candidate object and the secondary collision candidate object, and adds the calculated secondary collision path evaluation value and the value of the evaluation storage variable. The addition result is stored in the evaluation storage variable (step S2304).
- the secondary collision path evaluation value is (secondary collision possibility direction evaluation value) ⁇ (secondary collision possibility distance evaluation value) ⁇ (danger attribute combination evaluation value).
- step S2304 After the secondary collision path evaluation value calculation process (step S2304), the process returns to the secondary collision candidate object determination process (step S1203), and the same process is repeated for the remaining secondary collision candidate objects.
- step S1203 If there are no secondary collision candidate objects (NO in step S1203), the process returns to the primary collision candidate object determination process (step S1201), and the same process is repeated for the remaining primary collision candidate objects. It is. Every time the primary collision candidate object changes, all other objects are returned to the unacquired state as secondary collision candidate objects.
- the path evaluation unit 113e evaluates the value stored in the evaluation storage variable as a path evaluation.
- the value is stored in the evaluation holding unit 114 (step S1208).
- the action planning unit 115 acquires a route evaluation value from the evaluation holding unit 114, executes an action plan based on the acquired route evaluation value, and plans information indicating the planned action.
- the data is stored in the holding unit 116 (step S712).
- the damage risk can be reduced with respect to the primary collision and the secondary collision that occur when the robot moves on the movement path, and the effect is great.
- the gripping target object being transported (hereinafter referred to as “transporting target object”) is an object around the transport path.
- An action planning device that is supposed to collide with the user will be described. Specifically, for example, in FIG. 5, it is assumed that the robot holds the garlic at the position (40, 35, 0), moves in the X-axis direction, and carries the garlic to the position (70, 35, 0). Is done.
- FIG. 24 is a block diagram illustrating a configuration of an action planning apparatus 100f according to the eighth embodiment.
- the action planning device 100f replaces the gripping target information acquisition unit 105, the route evaluation unit 113, and the route generation unit 106 in the behavior planning device 100 with the gripping target information acquisition unit 105f, the route evaluation unit 113f, and the embodiment, respectively. 4 is used.
- the behavior planning apparatus 100f uses the secondary collision possibility evaluation unit 2103 during transportation without using the secondary collision possibility direction evaluation unit 108, the secondary collision possibility distance evaluation unit 109, and the robot position acquisition unit 101.
- the gripping target information acquisition unit 105f is a processing unit that acquires the position and type of the gripping target object to be transported.
- the gripping target information acquisition unit 105f acquires the position and type of the gripping target object by performing the same processing as the gripping target information acquisition unit 105 in the first embodiment.
- the gripping target information acquisition unit 105 f stores the acquired position of the gripping target object in the start point position holding unit 102. In this way, by setting the position where the robot grips the gripping target object as a starting point, it is possible to evaluate the risk of the gripping target object in the transport route. Specifically, for example, in FIG.
- the secondary collision possibility evaluation unit 2103 during transportation evaluates how easily the object in the specified environment is in a position where it collides with the movement path of the transportation target object.
- this evaluation is referred to as “secondary collision possibility evaluation during transportation”, and the evaluation value calculated at that time is referred to as “secondary collision possibility evaluation value during transportation”.
- the calculation method of the secondary collision possibility evaluation value during transportation is the same as the calculation method of the primary collision possibility evaluation value in the fourth embodiment. Specifically, for example, in FIG.
- the cake primary collision possibility evaluation value is .
- the distance from the line connecting the position (40, 35, 0) and the position (70, 35, 0) to the position (50, 15, 0) of the cake, and the value of the distance is “20”. Therefore, according to the same calculation method as the primary collision possibility evaluation value, the secondary collision possibility evaluation value during transportation is Euler number e to the -20th power (e -20 ).
- the route evaluation unit 113f is a processing unit that evaluates a risk of damage caused by contact or approach between the object to be transported and an object in the environment when the robot moves on the route held by the route holding unit 107. Damage risk is based on the evaluation of whether the object to be transported and the object in the environment are in a positional relationship that is easy to contact or approach and whether the object in the environment is a dangerous object with respect to the object to be transported. Evaluated.
- the object position attribute acquisition unit 103 acquires the object position attribute information of the object existing in the environment and stores it in the object position attribute holding unit 104 (step S702).
- the gripping target information acquisition unit 105f acquires the position of the gripping target object and outputs the position to the secondary collision possibility evaluation unit 2103 and the route evaluation unit 113f during transportation. In addition, the gripping target information acquisition unit 105 f outputs the attribute information of the gripping target object to the danger attribute combination evaluation unit 112. Furthermore, the gripping target information acquisition unit 105f stores the position of the gripping target object in the start point position holding unit 102 (step S703f). Specifically, for example, in FIG. 5, when the robot holds and transports garlic at the position (40, 35, 0), the coordinates of the start point position are (40, 35, 0).
- the route generation unit 106b acquires the destination of the robot from the outside, generates a route, and stores the generated route in the route holding unit 107 (step S2501). Specifically, for example, in FIG. 5, when the robot is designated to carry garlic to the position (70, 35, 0), the position (40, 35, 0) is set as the starting point, and the position (70, 35) is specified. , 0) becomes a route.
- the route evaluation unit 113f determines whether there is an object that is not acquired as a secondary collision candidate object among the objects other than the transport target object corresponding to the attribute information held by the object position attribute holding unit 104 (Ste S1203). If there is an unacquired object as a secondary collision candidate object (YES in step S1203), the path evaluation unit 113f arbitrarily selects one of the unacquired objects, and selects the selected object as a secondary object. Let it be a collision candidate object.
- the transportable secondary collision possibility evaluation unit 2103 calculates a secondary collision possibility evaluation value during transportation of the designated secondary collision candidate object with respect to the travel route of the transport target object (step S1202f).
- the moving path of the object to be transported is approximated by the moving path of the robot.
- the risk attribute combination evaluation unit 112 calculates a risk attribute combination evaluation value of the secondary collision candidate object and outputs it to the route evaluation unit 113f (step S709).
- the route evaluation unit 113f calculates the secondary collision possibility evaluation value during transportation calculated in the secondary collision possibility evaluation process during transportation (step S1202f) and the dangerous attribute combination calculated through the danger attribute combination acquisition process (step S709). Based on the evaluation value, the evaluation value of the secondary collision candidate object is calculated. That is, the damage risk when a secondary collision occurs is calculated (step S710).
- a specific calculation method is the same as that in the first embodiment. That is, the product of the secondary collision possibility evaluation value during transportation and the dangerous attribute combination evaluation value is taken as a damage risk when a secondary collision occurs.
- step S710 After the evaluation calculation process (step S710), the process returns to the secondary collision candidate object determination process (step S1203), and the same process is repeated for the remaining secondary collision candidate objects.
- the path evaluation unit 113f uses the path evaluation value based on the evaluation of each candidate calculated in the evaluation calculation process (step S710). Is calculated.
- the calculated route evaluation value is stored in the evaluation holding unit 114 (step S711).
- a specific calculation method is the same as that in the first embodiment.
- the action planning unit 115 acquires a route evaluation value from the evaluation holding unit 114, executes an action plan based on the acquired route evaluation value, and stores information indicating the planned action in the plan holding unit 116 (step). S712).
- a specific calculation method is the same as that in the first embodiment.
- Embodiment 8 As described above, according to the eighth embodiment, in a robot that moves in a home, when an object that is transported in contact with the mobile robot further contacts an object that is in the vicinity of the transport route, It is possible to determine whether or not damage will occur. Thereby, the action plan of a mobile robot can be performed. Along with this, secondary damage can be reduced.
- the route risk evaluation apparatus of the present invention has been described based on the embodiments, the present invention is not limited to these embodiments. Unless it deviates from the meaning of this invention, the form which carried out the various deformation
- the robot is exemplified as an example of the mobile device.
- the present invention can be applied to any mobile device that moves along a moving route such as an autonomous vehicle.
- the present invention can make an action plan for a mobile device such as a robot in consideration of the risk of secondary collision in a place where there are many objects and secondary collision occurs, such as a home environment.
- the present invention can be applied to an action plan device for making an action plan.
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Robotics (AREA)
- Mechanical Engineering (AREA)
- Manipulator (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
Description
本発明の実施の形態1では、指示された物体を把持するロボットの行動を計画する行動計画装置について説明する。行動計画装置は、ロボットが把持対象物体を把持する場合における、ロボットを把持対象物体に近づける際の移動速度を決定する。
ロボットが把持対象物体に接近する際、ロボットが把持対象物体を把持できずに、把持対象物体と衝突して、把持対象物体を弾いてしまう場合がある(以下、「1次衝突」と呼ぶ)。その原因は、取得した物体の位置の不正確さや、ロボットの制御の不正確さ、あるいは高速移動に対する要求に従ったロボットの高速移動等によるものである。また、ロボットに弾かれた物体が、その周辺物体に、さらに接触または接近(以下、「2次衝突」と呼ぶ)する場合がある。その際、ロボットと直接接触した第1物体と、第1物体と接触した第2物体との組合せによっては被害が発生する。
図1は、指示された物体を把持するロボットの行動を計画する行動計画装置100の構成を示すブロック図である。
では、2次衝突の可能性について説明する。ロボットと把持対象物体が1次衝突したとしても、常に2次衝突が発生するとは限らない。把持対象物体と周辺の物体が2次衝突するか否かの大まかな予想はシミュレーションによって取得することができる。
(A)2次衝突可能性方向評価部108より、2次衝突可能性方向評価を取得する。
(B)2次衝突可能性距離評価部109より、2次衝突可能性距離評価を取得する。
(C)危険属性組合せ評価部112より、危険属性組合せ評価を取得する。
(D)(2次衝突可能性方向評価)×(2次衝突可能性距離評価)×(危険属性組合せ評価)を指定された物体の被害リスクの評価とする。
行動計画装置100が行う行動計画処理の流れを図7に示す。
また、それぞれのベクトルの長さは、式2、式3となる。
実施の形態1において、ロボット位置取得部101は、GPSや画像処理を用いてロボット位置を取得する場合を示したが他の方法であっても構わない。例えば、ロボットの初期位置と、ロボットを初期位置から行動計画を開始する時点までに移動させた距離および方向とに基づいて、ロボットの位置を決定する構成としても構わない。
実施の形態1では、属性情報が物体の種別を示す情報である場合を示したが、「匂いが強い」、「磁力を発する」といった性質に関する情報であっても構わない。
実施の形態1では、属性情報が物体の種別を示す情報である場合を示したが、「“磁気”に弱い」、「“匂い”の強いものに弱い」といった情報であっても構わない。
実施の形態1では、簡単のため、把持対象物体の大きさについて考慮しなかったため、物体を把持する際、把持対象物体の位置まで移動する経路を計画した。しかし、実際には、物体には大きさがあるので、その大きさを考慮して、把持対象物体の若干手前で移動を停止する計画を立てるようにしても良い。
実施の形態1では、ロボットと把持対象物体との間の経路を考える際、2物体間に障害物がある場合や、2物体が同一平面上にない場合について言及しなかった。
実施の形態1における経路評価部113は、環境内にある物体を2次衝突の候補として、それぞれの候補に対して被害リスクを算出し、その合計値を経路評価値とした。しかし、被害リスクの最大値を経路評価値としても構わない。
実施の形態1における2次衝突可能性方向評価部108および2次衝突可能性距離評価部109は、物体の大きさを考慮することなく、2次衝突可能性方向評価値および2次衝突可能性距離評価値をそれぞれ算出する構成とした。しかし、物体の大きさを考慮した式を用いて2次衝突可能性方向評価値および2次衝突可能性距離評価値を算出する構成であっても構わない。なぜならば、把持対象物体から同じ距離に2つの物体が存在する場合、大きさが大きい物体ほど、2次衝突を起こす可能性が高いからである。なお、物体の大きさは、物体位置属性保持部104に記憶されているようにしてもよい。
実施の形態1における2次衝突可能性方向評価部108および2次衝突可能性距離評価部109は、周辺物体の大きさを考慮することなく、2次衝突可能性方向評価値および2次衝突可能性距離評価値をそれぞれ算出する構成とした。しかし、2次衝突可能性方向評価部108および2次衝突可能性距離評価部109は、周辺物体の大きさを考慮した式を用いて2次衝突可能性方向評価値および2次衝突可能性距離評価値を算出する構成であっても構わない。なぜならば、把持対象物体から同じ距離に2つの物体が存在する場合、大きさが大きい物体ほど、2次衝突を起こす可能性が高いからである。なお、物体の大きさは、物体位置属性保持部104に記憶されているようにしてもよい。
実施の形態1における2次衝突可能性方向評価部108および2次衝突可能性距離評価部109は、把持対象物体の大きさおよび周辺物体の大きさを考慮することなく、2次衝突可能性方向評価値および2次衝突可能性距離評価値をそれぞれ算出する構成とした。しかし、2次衝突可能性方向評価部108および2次衝突可能性距離評価部109は、両物体の大きさを考慮した式を用いて2次衝突可能性方向評価値および2次衝突可能性距離評価値を算出する構成であっても構わない。なぜならば、把持対象物体から同じ距離に2つの物体が存在する場合には、大きさが大きい物体ほど2次衝突を起こす可能性が高く、さらに、把持対象物体の大きさが大きいほど2次衝突を起こす可能性が高いからである。なお、物体の大きさは、物体位置属性保持部104に記憶されているようにしてもよい。
以上説明したように、実施の形態1によると、家庭内で移動するロボットにおいて、移動ロボットが接触または接近した物体による一次被害に加えて、当該物体がさらにその周辺物体と接触した場合に二次被害が発生するか否かも判断することができる。これにより、移動ロボットの行動計画を行うことができる。これに伴い、二次被害を低減することができる。
実施の形態2では、実施の形態1の行動計画装置100に基づき行動を行うロボットシステム805について説明する。
実施の形態2によると、2次衝突の被害リスクがある場合にロボットを低速で移動させることができる。よって、ロボットが把持対象物体にたどり着くまでの時間は、高速で移動する場合よりも余計にかかる。このため、ロボットが2次衝突をしそうな危険な経路を移動している場合であっても、そのことにユーザが気付き、行動変更指示装置802によって、ロボットを停止させることができる。
実施の形態3では、ロボットアームの行動計画を行う行動計画装置100aについて説明する。
実施の形態3の構成を用いれば、複数関節を有するロボットにおいても、2次衝突による被害を回避する行動計画を行うことができる。
実施の形態1では、ロボットと把持対象物体との1次衝突に基づき、2次衝突が発生することを想定してロボットの行動計画を立案する行動計画装置について説明した。しかしながら、ロボットは、把持対象物体以外にも、ロボットの移動経路上の物体と1次衝突する場合がある。その理由は、把持対象物体との衝突と同じく、家庭環境内でのロボットの位置認識精度が低いためであったり、ロボットの位置制御精度が低いためであったり、タスクの実行時間に対する制限が厳しいためである。
実施の形態1では、把持対象物体とロボットが1次衝突し、2次衝突が発生する場合の行動計画について述べた。ロボットが移動する場合には、移動経路上に存在する物体とロボットが1次衝突する可能性がある。さらに、ロボットの位置認識や制御に不正確さがある場合には、ロボットが移動経路上に存在していないと認識している物体とも1次衝突する。そこで、実施の形態4における行動計画装置100bは、環境内の物体をロボットとの1次衝突の候補とし、ロボットの移動経路からの距離が近い(とロボットが認識している)物体ほど、ロボットとの1次衝突が発生しやすいと評価する。また、距離が遠い物体ほど、ロボットとの1次衝突が発生しにくいと評価する。そして、各候補に対して、さらに、環境内のその他の物体を2次衝突の候補として、2次衝突が発生しやすいか否かと、発生した場合の被害度合いの大きさとを評価する。
図14は、実施の形態4に係る行動計画装置100bの構成を示すブロック図である。
行動計画装置100bが行う行動計画処理の流れを図15に示す。
実施の形態4では、ロボットの移動経路上の第1物体と、その周辺の第2物体とが2次衝突する被害リスクを算出することによる経路の危険性の評価方法を説明した。さらに、実施の形態4では、被害リスクに基づいて、第1物体付近でのロボットの移動速度を計画する行動計画について説明した。
実施の形態4の構成では、ロボットの移動経路上の物体とロボットが衝突することによって発生する1次衝突に対しても、被害リスクを低減することができ、その効果は大きい。
実施の形態4では、2次衝突の被害リスクが高い場合に低速で移動する行動を計画する行動計画装置100bを示した。しかし、低速で移動するだけでは2次衝突を回避できない場合がある。実施の形態5では、2次衝突の被害リスクが高い場合に、別の移動経路を計画する行動計画装置100cについて説明する。
実施の形態5における行動計画装置100cは、ロボットが移動先に移動する経路を複数の経路の集まりとして計画する。図16は、位置(10,35,0)にあるロボット401が、位置(40,35,0)の移動先1304に移動する際、ロボット401の移動経路を、3つの経路(経路1301、1302および1303)からなる経路群として計画した場合の一例を示す例である。
図18は、実施の形態5に係る行動計画装置100cの構成を示すブロック図である。
実施の形態5における経路生成部106cは、ロボットの位置から移動先までをはじめから複数の経路として経路設計したが一つの直線の経路とすることができる部分であれば、複数の経路としなくても構わない。例えば、一つの直線として被害リスクを算出し、被害リスクが所定の値よりも低ければその直線を経路とし、被害リスクが所定の値よりも高ければ、経路を細分化して、どの部分の経路の被害リスクが高いかを見つけ出し、その部分を除外して改良経路を設計しても構わない。このような、構成を用いれば、被害リスクを算出する計算負荷を減らすことができる。
実施の形態5における行動計画部115cは、経路の評価が悪かった場合、経路の始点と終点を直径とする円(または球)の領域を除外領域としたが、他の形状であっても構わない。つまり、被害リスクが高い経路が多い場合に、除外する領域を大きめに取れば、改良経路を何度も作り直すことを防ぐことができることが考えられる。
実施の形態5では、実施の形態4に示した行動計画装置100bに対して、別経路を計画するようにしたが、実施の形態1に示した行動計画装置100に対して、別経路を計画するようにしてもよい。つまり、1次衝突可能性評価部1101が設けられておらず、2次衝突可能性方向評価部108b、2次衝突可能性距離評価部109bおよび危険属性組合せ評価部112bの代わりに、2次衝突可能性方向評価部108、2次衝突可能性距離評価部109および危険属性組合せ評価部112を用いた構成であってもよい。
実施の形態5の構成では、ロボットの移動経路を変更することで被害リスクを低減することができ、その効果は大きい。
実施の形態4では、2次衝突の被害リスクが高い場合に低速で移動する行動を計画する行動計画装置100bを示した。しかし、ロボットがタスクを実行する際に、複数の移動経路が想定される場合がある。実施の形態6では、複数の行動計画のうち、2次衝突の被害リスクが低い行動を選択する行動計画装置100dについて説明する。
実施の形態6では、同じ名称の物体が複数あり、複数の移動経路がある場合に、その中から被害リスクが小さい経路を選択する場合について説明したが、これに限られる必要はない。例えば、1つの物体に把持ポイントが複数ある場合に、それぞれの把持ポイントに対する経路のうち、最も被害リスクが小さいものを選択する構成であっても構わない。
実施の形態6における行動計画部115dは、被害リスクが最も少ない経路を選択する構成としたが、さらに別の要素を考慮して経路選択を行う構成としても構わない。例えば、移動経路長が長い経路はエネルギーや時間といったコストを余計に消費するので、被害リスクと経路長とを組み合わせて、経路評価を行う構成としても構わない。
実施の形態6の構成では、ロボットが指定された目的を達成するための移動経路が複数ある場合に、より2次衝突の被害リスクが少ない経路を選択することができる。よって、ロボットによる目的達成の際に生じる被害リスクを低減することができ、その効果は大きい。
実施の形態4では、2次衝突による被害リスクに基づき、ロボットに対する移動経路の危険性を評価した。しかし、実際にロボットがある移動経路を移動する際には、1次衝突による被害リスクも存在する。
図21は、実施の形態7に係る行動計画装置100eの構成を示すブロック図である。
行動計画装置100eが行う行動計画処理の流れを図23に示す。
実施の形態7の構成では、ロボットが移動経路上を移動することによって発生する1次衝突および2次衝突に対して、被害リスクを低減することができ、その効果は大きい。
実施の形態1では、ロボットが把持対象物体と1次衝突し、把持対象物体がさらに周辺物体と2次衝突することを想定した行動計画装置について説明した。
ロボットが運搬対象物体を運搬している際、ロボットが運搬対象物体を落としたり、弾き飛ばしてしまう場合がある。その原因は、運搬対象物体を落とさないように握るのに必要な力の推定の不正確さや、地面の凹凸やロボット自体の動力に起因する振動、スリップによる転倒等によるものである。ロボットが落とした運搬対象物体、ロボットが弾き飛ばした運搬対象物体が、その周辺物体に、2次衝突する場合がある。その際、ロボットにより運搬される第1物体と、第1物体と接触した第2物体との組合せによっては被害が発生する。
図24は、実施の形態8に係る行動計画装置100fの構成を示すブロック図である。以下、行動計画装置100fと実施の形態1の行動計画装置100との違いについて説明する。つまり、行動計画装置100fは、行動計画装置100における把持対象情報取得部105、経路評価部113、経路生成部106の代わりに、それぞれ、把持対象情報取得部105f、経路評価部113f、実施の形態4に記載の経路生成部106bを用いる。
行動計画装置100fが行う行動計画処理の流れを図25に示す。
以上説明したように、実施の形態8によると、家庭内で移動するロボットにおいて、移動ロボットに接触して運搬されている物体が、さらにその運搬経路の周辺にある物体と接触した場合に二次被害が発生するか否かを判断することができる。これにより、移動ロボットの行動計画を行うことができる。これに伴い、二次被害を低減することができる。
101、101a ロボット位置取得部
102、102a 始点位置保持部
103 物体位置属性取得部
104 物体位置属性保持部
105、105f 把持対象情報取得部
106、106a、106b、106c、106d 経路生成部
107、107a、107c 経路保持部
108、108b 2次衝突可能性方向評価部
109、109b 2次衝突可能性距離評価部
111 危険属性組合せ情報保持部
112、112b 危険属性組合せ評価部
113、113a、113b、113e、113f 経路評価部
114、114c 評価保持部
115、115c、115d 行動計画部
116、116c、116d 計画保持部
401、800 ロボット
801 移動装置
802 行動変更指示装置
803 作業台
805 ロボットシステム
901 ロボットアーム
1101 1次衝突可能性評価部
1501 除外領域保持部
2101 1次衝突危険属性情報保持部
2102 1次衝突危険属性評価部
2103 運搬時2次衝突可能性評価部
Claims (17)
- 移動機器の移動経路の危険性を評価する経路危険性評価装置であって、
移動機器の移動経路を示す情報である移動経路情報を保持する経路保持部と、
前記移動機器が移動する環境内に存在する複数の物体の位置を示す位置情報を保持する物体位置属性保持部と、
前記環境内に存在する物体同士が接触または接近した際の被害度合いを示す情報を保持する危険属性組合せ情報保持部と、
前記経路保持部に保持されている移動経路情報で示される前記移動機器の移動経路と、前記物体位置属性保持部に保持されている位置情報で示される前記複数の物体の位置との関係に基づいて、前記移動機器と接触することにより移動させられる可能性がある第1物体が、さらに第2物体に接触または接近する可能性を評価する2次衝突可能性評価部と、
前記危険属性組合せ情報保持部に保持されている前記被害度合いを示す情報を参照することにより、前記第1物体と前記第2物体とが接触または接近した際の被害度合いを決定する危険属性組合せ評価部と、
前記2次衝突可能性評価部による評価結果と、前記危険属性組合せ評価部により決定された被害度合いとに基づいて、前記移動機器に対する前記移動経路の危険性を評価する経路評価部と
を備える経路危険性評価装置。 - 前記経路評価部は、前記2次衝突可能性評価部における評価を示す評価値と、前記危険属性組合せ評価部により決定された前記被害度合いを示す値とを乗算した乗算値が大きいほど、前記移動機器に対する前記移動経路の危険性が高いと評価する
請求項1記載の経路危険性評価装置。 - 前記2次衝突可能性評価部は、前記移動経路における前記移動機器の移動方向と前記第1物体の位置および前記第2物体の位置を結ぶ直線とがなす角度が0度に近づくほど大きくなる2次衝突可能性方向評価値を算出することにより、前記第1物体が前記第2物体に接触または接近する可能性を評価する2次衝突可能性方向評価部を含む
請求項1または2に記載の経路危険性評価装置。 - 前記2次衝突可能性方向評価部は、前記移動経路を平行移動させることで得られる、前記第1物体の中心位置を通る直線が前記第2物体の外接円と交差する場合に最大となる2次衝突可能性方向評価値を算出することにより、前記第1物体が前記第2物体に接触または接近する可能性を評価する
請求項3記載の経路危険性評価装置。 - 前記2次衝突可能性方向評価部は、前記移動経路を平行移動させることで得られる、前記第1物体の中心位置を通る直線上を前記第1物体の外接円が通過すると仮定したときに、前記第1物体の外接円と前記第2物体の外接円とが接触する場合に最大となる2次衝突可能性方向評価値を算出することにより、前記第1物体が前記第2物体に接触または接近する可能性を評価する
請求項3記載の経路危険性評価装置。 - 前記2次衝突可能性評価部は、前記第1物体と前記第2物体との間の距離が小さいほど大きくなる2次衝突可能性距離評価値を算出することにより、前記第1物体が前記第2物体に接触または接近する可能性を評価する2次衝突可能性距離評価部を含む
請求項1~5のいずれか1項に記載の経路危険性評価装置。 - 前記2次衝突可能性距離評価部は、前記第1物体と前記第2物体との間の距離が小さく、かつ前記第2物体の大きさが大きいほど大きくなる2次衝突可能性距離評価値を算出することにより、
前記第1物体が前記第2物体に接触または接近する可能性を評価する
請求項6記載の経路危険性評価装置。 - さらに、前記移動経路と前記第1物体との間の距離が小さいほど大きくなる1次衝突可能性評価値を算出することにより、前記移動機器が前記第1物体に接触または接近する可能性を示す評価する1次衝突可能性評価部を備え、
前記経路評価部は、前記2次衝突可能性評価部による評価結果と、前記危険属性組合せ評価部により決定された被害度合いと、前記1次衝突可能性評価部による評価結果とに基づいて、前記移動機器に対する前記移動経路の危険性を評価する
請求項1~7のいずれか1項に記載の経路危険性評価装置。 - さらに、
前記移動機器と前記環境内に存在する物体とが接触または接近した際の被害度合いを示す情報を保持する1次衝突危険属性情報保持部と、
前記1次衝突危険属性情報保持部に保持されている前記被害度合いを示す情報を参照することにより、前記移動機器と前記第1物体とが接触または接近した際の被害度合いを決定する危険属性組合せ評価部とを備え、
前記経路評価部は、前記2次衝突可能性評価部による評価結果と、前記危険属性組合せ評価部により決定された被害度合いと、前記1次衝突可能性評価部による評価結果と、前記危険属性組合せ評価部により決定された被害度合いとに基づいて、前記移動機器に対する前記移動経路の危険性を評価する
請求項8記載の経路危険性評価装置。 - 前記経路保持部は、前記移動機器における複数箇所の移動経路を示す情報である移動経路情報を保持し、
前記2衝突可能性評価部は、前記移動機器における前記複数箇所の各々について、前記第1物体が前記第2物体に接触または接近する可能性を評価し、
前記経路評価部は、前記移動機器の各箇所に対する前記移動経路の危険性を評価し、前記複数箇所について、各箇所における評価結果を集計することにより、前記移動機器に対する前記移動経路の危険性を評価する
請求項1~9のいずれか1項に記載の経路危険性評価装置。 - さらに、前記経路評価部による評価結果に従い、前記移動機器に対する前記移動経路の危険性が高いほど前記移動経路における前記移動機器の移動速度を遅い速度に決定する行動計画部を備える
請求項1~10のいずれか1項に記載の経路危険性評価装置。 - 前記第1物体は、前記移動機器により操作が施される対象の物体とは異なる物体であり、
前記経路危険性評価装置は、さらに、前記経路評価部により評価された前記移動経路の危険性が所定の危険性を超えるか否かを判断し、超える場合には、前記移動経路を通らないような前記移動機器の移動経路を生成し、前記経路保持部に保持する経路生成部を備える
請求項1~11のいずれか1項に記載の経路危険性評価装置。 - 前記第1物体は、前記移動機器により操作が施される対象の物体であり、
前記経路危険性評価装置は、さらに、前記第1物体が複数存在する場合には、複数の第1物体までの移動経路のうち、前記経路評価部によって最も危険性が低いと判断される移動経路を、前記移動機器の移動経路と決定する行動計画部を備える
請求項1~10のいずれか1項に記載の経路危険性評価装置。 - 前記危険属性組合せ情報保持部は、前記環境内に存在する物体同士が接触または接近した際の被害金額を保持し、
前記危険属性組合せ評価部は、前記危険属性組合せ情報保持部に保持されている前記被害金額に基づいて、前記第1物体と前記第2物体とが接触または接近した際の被害金額を、前記第1物体と前記第2物体とが接触または接近した際の被害度合いとして決定する
請求項1~13のいずれか1項に記載の経路危険性評価装置。 - 前記第1物体は、前記移動機器により運搬させられる物体である
請求項1~11のいずれか1項に記載の経路危険性評価装置。 - 移動機器の移動経路の危険性を評価する経路危険性評価方法であって、
経路保持部に保持されている移動経路情報で示される前記移動機器の移動経路と、物体位置属性保持部に保持されている位置情報で示される前記移動機器が移動する環境内に存在する複数の物体の位置との関係に基づいて、前記移動機器と接触または接近することにより移動させられる第1物体が、さらに第2物体に接触または接近する可能性を評価する可能性評価ステップと、
危険属性組合せ情報保持部に保持されている前記環境内に存在する物体同士が接触または接近した際の被害度合いを示す情報を参照することにより、前記第1物体と前記第2物体とが接触または接近した際の被害度合いを決定する決定ステップと、
前記可能性評価ステップでの評価結果と、前記決定ステップで決定された被害度合いとに基づいて、前記移動機器に対する前記移動経路の危険性を評価する危険性評価ステップと
を含む経路危険性評価方法。 - 移動機器の移動経路の危険性を評価するコンピュータ実行可能なプログラムであって、
経路保持部に保持されている移動経路情報で示される前記移動機器の移動経路と、物体位置属性保持部に保持されている位置情報で示される前記移動機器が移動する環境内に存在する複数の物体の位置との関係に基づいて、前記移動機器と接触または接近することにより移動させられる第1物体が、さらに第2物体に接触または接近する可能性を評価する可能性評価ステップと、
危険属性組合せ情報保持部に保持されている前記環境内に存在する物体同士が接触または接近した際の被害度合いを示す情報を参照することにより、前記第1物体と前記第2物体とが接触または接近した際の被害度合いを決定する決定ステップと、
前記可能性評価ステップでの評価結果と、前記決定ステップで決定された被害度合いとに基づいて、前記移動機器に対する前記移動経路の危険性を評価する危険性評価ステップと
をコンピュータに実行させるためのプログラム。
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US12/671,144 US7844398B2 (en) | 2008-07-09 | 2009-07-08 | Path risk evaluating apparatus |
| CN2009801065246A CN101959653B (zh) | 2008-07-09 | 2009-07-08 | 路径危险性评价装置 |
| JP2009549332A JP4495252B2 (ja) | 2008-07-09 | 2009-07-08 | 経路危険性評価装置、経路危険性評価方法及びプログラム |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2008179564 | 2008-07-09 | ||
| JP2008-179564 | 2008-07-09 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2010004744A1 true WO2010004744A1 (ja) | 2010-01-14 |
Family
ID=41506869
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2009/003183 Ceased WO2010004744A1 (ja) | 2008-07-09 | 2009-07-08 | 経路危険性評価装置 |
Country Status (4)
| Country | Link |
|---|---|
| US (1) | US7844398B2 (ja) |
| JP (1) | JP4495252B2 (ja) |
| CN (1) | CN101959653B (ja) |
| WO (1) | WO2010004744A1 (ja) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2017010234A (ja) * | 2015-06-22 | 2017-01-12 | 株式会社日立製作所 | 移動ロボットシステム及び移動ロボットの制御方法 |
| CN108778636A (zh) * | 2016-02-08 | 2018-11-09 | 伯克希尔格雷股份有限公司 | 用于提供使用运动规划的各种物体的处理的系统和方法 |
| JP2019206080A (ja) * | 2017-08-18 | 2019-12-05 | 三菱電機株式会社 | ロボット制御装置およびこれを用いたロボットシステム |
| WO2020012944A1 (ja) * | 2018-07-09 | 2020-01-16 | ソニー株式会社 | 制御装置、制御方法、およびプログラム |
| CN114494848A (zh) * | 2021-12-21 | 2022-05-13 | 重庆特斯联智慧科技股份有限公司 | 一种机器人视距路径确定方法和装置 |
| WO2023047742A1 (ja) * | 2021-09-22 | 2023-03-30 | パナソニックIpマネジメント株式会社 | 移動ロボット制御システム、及び移動ロボット制御方法 |
| US12440984B2 (en) | 2020-05-25 | 2025-10-14 | Fanuc Corporation | Offline teaching device and motion-program generation method |
Families Citing this family (58)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090105939A1 (en) * | 2007-10-22 | 2009-04-23 | Toyota Motor Engineering & Manufacturing North America, Inc. | Vehicle navigation system with obstacle avoidance |
| KR101667029B1 (ko) * | 2009-08-10 | 2016-10-17 | 삼성전자 주식회사 | 로봇의 경로 계획방법 및 장치 |
| DE112011103155T5 (de) * | 2010-09-21 | 2013-07-18 | Toyota Jidosha Kabushiki Kaisha | Beweglicher Körper |
| DE102010063208A1 (de) * | 2010-12-16 | 2012-06-21 | Robert Bosch Gmbh | Verfahren zum Betreiben einer Sicherungseinrichtung für eine Handhabungsvorrichtung, Sicherungseinrichtung für eine Handhabungsvorrichtung und Handhabungsvorrichtung |
| US9230240B2 (en) * | 2011-01-21 | 2016-01-05 | Qualcomm Incorporated | Contact relevance based on context |
| US9160898B2 (en) | 2011-01-25 | 2015-10-13 | Autofuss | System and method for improved video motion control |
| CN102346483B (zh) * | 2011-06-07 | 2012-11-28 | 南京航空航天大学 | 基于无源rfid和辅助视觉的agv运行控制方法 |
| JP6081061B2 (ja) * | 2011-08-04 | 2017-02-15 | オリンパス株式会社 | 手術支援装置 |
| JP6005950B2 (ja) | 2011-08-04 | 2016-10-12 | オリンパス株式会社 | 手術支援装置及びその制御方法 |
| JP5936914B2 (ja) | 2011-08-04 | 2016-06-22 | オリンパス株式会社 | 操作入力装置およびこれを備えるマニピュレータシステム |
| WO2013018861A1 (ja) | 2011-08-04 | 2013-02-07 | オリンパス株式会社 | 医療用マニピュレータおよびその制御方法 |
| US9161772B2 (en) | 2011-08-04 | 2015-10-20 | Olympus Corporation | Surgical instrument and medical manipulator |
| JP6000641B2 (ja) | 2011-08-04 | 2016-10-05 | オリンパス株式会社 | マニピュレータシステム |
| JP6021484B2 (ja) | 2011-08-04 | 2016-11-09 | オリンパス株式会社 | 医療用マニピュレータ |
| WO2013018908A1 (ja) | 2011-08-04 | 2013-02-07 | オリンパス株式会社 | 医療用マニピュレータおよび手術支援装置 |
| JP5953058B2 (ja) | 2011-08-04 | 2016-07-13 | オリンパス株式会社 | 手術支援装置およびその着脱方法 |
| JP5931497B2 (ja) | 2011-08-04 | 2016-06-08 | オリンパス株式会社 | 手術支援装置およびその組立方法 |
| JP6021353B2 (ja) | 2011-08-04 | 2016-11-09 | オリンパス株式会社 | 手術支援装置 |
| JP6009840B2 (ja) | 2011-08-04 | 2016-10-19 | オリンパス株式会社 | 医療機器 |
| JP5841451B2 (ja) | 2011-08-04 | 2016-01-13 | オリンパス株式会社 | 手術器具およびその制御方法 |
| US9832352B2 (en) | 2011-11-16 | 2017-11-28 | Autofuss | System and method for 3D projection mapping with robotically controlled objects |
| JP2015506121A (ja) * | 2011-11-16 | 2015-02-26 | ストーウェル・リンネル・エルエルシー | ロボット制御のオブジェクトによる3dプロジェクション・マッピングのシステム及び方法 |
| KR101800189B1 (ko) * | 2012-04-30 | 2017-11-23 | 삼성전자주식회사 | 수술 로봇의 힘 제어 장치 및 방법 |
| US9176500B1 (en) | 2012-05-14 | 2015-11-03 | Google Inc. | Consideration of risks in active sensing for an autonomous vehicle |
| US9043140B2 (en) | 2012-06-29 | 2015-05-26 | Here Global B.V. | Predictive natural guidance |
| US9043070B2 (en) | 2012-08-22 | 2015-05-26 | Industrial Technology Research Institute | Moving device and moving control method thereof |
| TWM451103U (zh) * | 2012-10-30 | 2013-04-21 | Agait Technology Corp | 行走裝置 |
| CN104798123B (zh) * | 2012-11-21 | 2016-11-02 | 丰田自动车株式会社 | 驾驶辅助装置以及驾驶辅助方法 |
| US8843303B1 (en) * | 2012-12-17 | 2014-09-23 | Rockwell Collins, Inc. | Risk-aware contingency flight re-planner system and related method |
| US8972057B1 (en) * | 2013-01-09 | 2015-03-03 | The Boeing Company | Systems and methods for generating a robotic path plan in a confined configuration space |
| KR20240027152A (ko) * | 2014-02-20 | 2024-02-29 | 엠비엘 리미티드 | 로봇식 요리용 키친에서의 음식 준비를 위한 방법 및 시스템 |
| KR101664582B1 (ko) * | 2014-11-12 | 2016-10-10 | 현대자동차주식회사 | 자율주행차량의 주행경로 생성장치 및 방법 |
| US9926138B1 (en) * | 2015-09-29 | 2018-03-27 | Amazon Technologies, Inc. | Determination of removal strategies |
| US10924881B2 (en) * | 2016-03-03 | 2021-02-16 | Husqvarna Ab | Device for determining construction device and worker position |
| CN109070366B (zh) * | 2016-05-16 | 2021-09-28 | 三菱电机株式会社 | 机器人动作评价装置、机器人动作评价方法及机器人系统 |
| TWI641856B (zh) * | 2016-05-24 | 2018-11-21 | 華碩電腦股份有限公司 | 移動裝置以及移動控制方法 |
| JP6672076B2 (ja) * | 2016-05-27 | 2020-03-25 | 株式会社東芝 | 情報処理装置及び移動体装置 |
| US20180068392A1 (en) * | 2016-09-02 | 2018-03-08 | Allstate Insurance Company | Generating an Output Based on Processed Sensor Data |
| TWI615691B (zh) * | 2016-11-24 | 2018-02-21 | 財團法人資訊工業策進會 | 防碰撞系統及防碰撞方法 |
| CN110049934B (zh) | 2016-12-09 | 2022-02-25 | 伯克希尔格雷股份有限公司 | 用于处理在车辆中设置的物品的系统和方法 |
| US10475343B2 (en) * | 2017-05-18 | 2019-11-12 | Ford Global Technologies, Llc | Vehicle ice strike prediction |
| KR101976358B1 (ko) * | 2017-10-31 | 2019-08-28 | 경희대학교 산학협력단 | 로봇의 안전성 향상 방법 |
| WO2019088681A1 (ko) | 2017-10-31 | 2019-05-09 | 경희대학교산학협력단 | 로봇의 안전성 향상 및 안전성 평가 방법 |
| US11119216B1 (en) | 2017-11-02 | 2021-09-14 | AI Incorporated | Efficient coverage planning of mobile robotic devices |
| US12038756B2 (en) * | 2017-12-19 | 2024-07-16 | Carnegie Mellon University | Intelligent cleaning robot |
| EP3518059B1 (de) * | 2018-01-24 | 2020-04-01 | Siemens Aktiengesellschaft | Verfahren zur rechnergestützten benutzerassistenz bei der in-betriebnahme eines bewegungsplaners für eine maschine |
| JP6897597B2 (ja) * | 2018-02-16 | 2021-06-30 | トヨタ自動車株式会社 | 駐車支援装置 |
| KR102749960B1 (ko) * | 2018-12-26 | 2025-01-03 | 삼성전자주식회사 | 청소 로봇 및 그의 태스크 수행 방법 |
| KR102715879B1 (ko) * | 2019-03-07 | 2024-10-14 | 삼성전자주식회사 | 전자 장치 및 그 제어 방법 |
| US11440183B2 (en) * | 2019-03-27 | 2022-09-13 | Abb Schweiz Ag | Hybrid machine learning-based systems and methods for training an object picking robot with real and simulated performance data |
| WO2020242065A1 (ko) * | 2019-05-31 | 2020-12-03 | 주식회사 라운지랩 | 위험도 판단에 기초한 로봇 움직임 제어 방법 및 이를 이용한 이동 로봇 장치 |
| AT523218B1 (de) * | 2019-12-13 | 2021-12-15 | Tgw Logistics Group Gmbh | Lager- und Kommissioniersystem und Verfahren zum Prognostizieren und/oder Abwenden einer zukünftigen Störung |
| US11634123B2 (en) | 2020-07-09 | 2023-04-25 | Toyota Research Institute, Inc. | Methods and systems for prioritizing computing methods for autonomous vehicles |
| CN113561186B (zh) * | 2021-09-24 | 2021-12-07 | 季华实验室 | 一种机械手路径规划结果的评价方法 |
| US11866269B2 (en) | 2021-10-06 | 2024-01-09 | Berkshire Grey Operating Company, Inc. | Dynamic processing of objects provided in elevated vehicles with evacuation systems and methods for receiving objects |
| WO2025033944A1 (ko) * | 2023-08-08 | 2025-02-13 | 주식회사 세이프틱스 | 로봇의 안전성 평가 및 안전성 향상 방법 |
| WO2025033942A1 (ko) * | 2023-08-08 | 2025-02-13 | 주식회사 세이프틱스 | 로봇의 안전성 평가 및 안전성 향상 방법 |
| CN117621103B (zh) * | 2023-12-22 | 2024-08-23 | 沈阳工业大学 | 基于位姿自适应逆透视变换的护理机器人测距及避障方法 |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH03118610A (ja) * | 1989-09-29 | 1991-05-21 | Mazda Motor Corp | 移動車の走行制御装置 |
| WO2006043396A1 (ja) * | 2004-10-19 | 2006-04-27 | Matsushita Electric Industrial Co., Ltd. | ロボット装置 |
| JP2008137127A (ja) * | 2006-12-04 | 2008-06-19 | Matsushita Electric Ind Co Ltd | ロボットの制御方法 |
Family Cites Families (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CH650738A5 (de) * | 1980-11-14 | 1985-08-15 | Inventio Ag | Einrichtung zur abstandhaltung von spurgebundenen fahrzeugen. |
| JPS61105594A (ja) | 1984-10-29 | 1986-05-23 | セイコーエプソン株式会社 | 液晶表示制御回路 |
| JPS6385906A (ja) | 1986-09-30 | 1988-04-16 | Toshiba Corp | ロボツトの環境モデル構築システム |
| JPH0772925A (ja) | 1993-09-06 | 1995-03-17 | Yamatake Honeywell Co Ltd | 障害物回避制御装置 |
| US5835684A (en) * | 1994-11-09 | 1998-11-10 | Amada Company, Ltd. | Method for planning/controlling robot motion |
| US6129025A (en) * | 1995-07-04 | 2000-10-10 | Minakami; Hiroyuki | Traffic/transportation system |
| JPH0970786A (ja) | 1995-09-06 | 1997-03-18 | Ricoh Co Ltd | 把持ハンド |
| JP3781557B2 (ja) * | 1998-06-24 | 2006-05-31 | 本田技研工業株式会社 | 物品の最適移送経路決定方法 |
| EP1233387A2 (en) * | 2001-02-19 | 2002-08-21 | Hitachi Kokusai Electric Inc. | Vehicle emergency reporting system and method |
| JP3861818B2 (ja) | 2003-01-28 | 2006-12-27 | 株式会社日立製作所 | 車間距離制御装置 |
| DE10327600A1 (de) * | 2003-06-18 | 2005-01-27 | Siemens Ag | Vorrichtung und Verfahren zur Kollisionsüberwachung einer Maschinenkomponente mit einem Werkstück oder anderen Maschinenkomponenten |
| KR20050020328A (ko) * | 2003-08-22 | 2005-03-04 | 삼성전자주식회사 | 습식 세정설비의 웨이퍼 이송시스템 |
| JP3930490B2 (ja) | 2004-04-23 | 2007-06-13 | ファナック株式会社 | 物品取出し装置 |
| US20060058921A1 (en) | 2004-09-13 | 2006-03-16 | Tamao Okamoto | Mobile robot |
| JP4464893B2 (ja) | 2004-09-13 | 2010-05-19 | パナソニック株式会社 | 移動ロボット |
| JP4621073B2 (ja) * | 2005-05-23 | 2011-01-26 | 本田技研工業株式会社 | ロボット制御装置 |
| JP5112666B2 (ja) | 2006-09-11 | 2013-01-09 | 株式会社日立製作所 | 移動装置 |
| JP4550849B2 (ja) * | 2007-03-22 | 2010-09-22 | 株式会社東芝 | アーム搭載移動ロボット |
| JP4661838B2 (ja) * | 2007-07-18 | 2011-03-30 | トヨタ自動車株式会社 | 経路計画装置及び方法、コスト評価装置、並びに移動体 |
-
2009
- 2009-07-08 JP JP2009549332A patent/JP4495252B2/ja active Active
- 2009-07-08 WO PCT/JP2009/003183 patent/WO2010004744A1/ja not_active Ceased
- 2009-07-08 US US12/671,144 patent/US7844398B2/en active Active
- 2009-07-08 CN CN2009801065246A patent/CN101959653B/zh active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH03118610A (ja) * | 1989-09-29 | 1991-05-21 | Mazda Motor Corp | 移動車の走行制御装置 |
| WO2006043396A1 (ja) * | 2004-10-19 | 2006-04-27 | Matsushita Electric Industrial Co., Ltd. | ロボット装置 |
| JP2008137127A (ja) * | 2006-12-04 | 2008-06-19 | Matsushita Electric Ind Co Ltd | ロボットの制御方法 |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2017010234A (ja) * | 2015-06-22 | 2017-01-12 | 株式会社日立製作所 | 移動ロボットシステム及び移動ロボットの制御方法 |
| CN108778636A (zh) * | 2016-02-08 | 2018-11-09 | 伯克希尔格雷股份有限公司 | 用于提供使用运动规划的各种物体的处理的系统和方法 |
| CN108778636B (zh) * | 2016-02-08 | 2021-11-19 | 伯克希尔格雷股份有限公司 | 用于提供使用运动规划的各种物体的处理的系统和方法 |
| JP2019206080A (ja) * | 2017-08-18 | 2019-12-05 | 三菱電機株式会社 | ロボット制御装置およびこれを用いたロボットシステム |
| WO2020012944A1 (ja) * | 2018-07-09 | 2020-01-16 | ソニー株式会社 | 制御装置、制御方法、およびプログラム |
| US12038750B2 (en) | 2018-07-09 | 2024-07-16 | Sony Corporation | Control device and control method for a movable body |
| US12440984B2 (en) | 2020-05-25 | 2025-10-14 | Fanuc Corporation | Offline teaching device and motion-program generation method |
| WO2023047742A1 (ja) * | 2021-09-22 | 2023-03-30 | パナソニックIpマネジメント株式会社 | 移動ロボット制御システム、及び移動ロボット制御方法 |
| CN114494848A (zh) * | 2021-12-21 | 2022-05-13 | 重庆特斯联智慧科技股份有限公司 | 一种机器人视距路径确定方法和装置 |
| CN114494848B (zh) * | 2021-12-21 | 2024-04-16 | 重庆特斯联智慧科技股份有限公司 | 一种机器人视距路径确定方法和装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| US7844398B2 (en) | 2010-11-30 |
| US20100217528A1 (en) | 2010-08-26 |
| CN101959653A (zh) | 2011-01-26 |
| CN101959653B (zh) | 2012-10-24 |
| JP4495252B2 (ja) | 2010-06-30 |
| JPWO2010004744A1 (ja) | 2011-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP4495252B2 (ja) | 経路危険性評価装置、経路危険性評価方法及びプログラム | |
| EP3650181B1 (en) | Route output method, route output system and route output program | |
| Mainprice et al. | Human-robot collaborative manipulation planning using early prediction of human motion | |
| Althoff et al. | Safety assessment of robot trajectories for navigation in uncertain and dynamic environments | |
| Boroushaki et al. | Robotic grasping of fully-occluded objects using rf perception | |
| US8280837B2 (en) | Contact state estimation for multi-finger robot hands using particle filters | |
| US20180056512A1 (en) | Trajectory Planning with Droppable Objects | |
| JP6032163B2 (ja) | 三次元物体認識装置、三次元物体認識方法、及び移動体 | |
| CN111226237A (zh) | 用于稳健抓取和瞄准物体的机器人系统和方法 | |
| Coltin et al. | Multi-observation sensor resetting localization with ambiguous landmarks | |
| WO2021095464A1 (ja) | ロボット制御モデル学習方法、ロボット制御モデル学習装置、ロボット制御モデル学習プログラム、ロボット制御方法、ロボット制御装置、ロボット制御プログラム、及びロボット | |
| US10933526B2 (en) | Method and robotic system for manipulating instruments | |
| Cimurs et al. | Information-based heuristics for learned goal-driven exploration and mapping | |
| Makita et al. | 3D two-fingered caging for two types of objects: sufficient conditions and planning | |
| CN117234208A (zh) | 一种室内物品运输机器人及其控制方法 | |
| JP6528641B2 (ja) | 自己位置推定方法 | |
| Lee et al. | Velocity range-based reward shaping technique for effective map-less navigation with LiDAR sensor and deep reinforcement learning | |
| Huo et al. | Rearranging Deformable Linear Objects for Implicit Goals with Self‐Supervised Planning and Control | |
| Tsai et al. | Cooperative localization using fuzzy decentralized extended information filtering for homogenous omnidirectional mobile multi-robot system | |
| Vatcha et al. | Practical motion planning in unknown and unpredictable environments | |
| JP2024507761A (ja) | 動的環境におけるロボットのナビゲーションをヒューリスティックな学習に基づき制御するシステムおよび方法 | |
| Mao et al. | Object shape estimation through touch-based continuum manipulation | |
| Saini et al. | Planning Robot Placement for Object Grasping | |
| CN120134330B (zh) | 双臂机器人的避障方法以及系统 | |
| Umam et al. | Optimal mapping trajectory for autonomous robot waiter |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| WWE | Wipo information: entry into national phase |
Ref document number: 200980106524.6 Country of ref document: CN |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2009549332 Country of ref document: JP |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 12671144 Country of ref document: US |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 09794192 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 09794192 Country of ref document: EP Kind code of ref document: A1 |