WO2019112407A1 - Navigation et comportement autonomes de véhicule sans pilote sans liaison avec la station de contrôle - Google Patents
Navigation et comportement autonomes de véhicule sans pilote sans liaison avec la station de contrôle Download PDFInfo
- Publication number
- WO2019112407A1 WO2019112407A1 PCT/MA2018/050013 MA2018050013W WO2019112407A1 WO 2019112407 A1 WO2019112407 A1 WO 2019112407A1 MA 2018050013 W MA2018050013 W MA 2018050013W WO 2019112407 A1 WO2019112407 A1 WO 2019112407A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- module
- mission
- behavior
- execution
- tree
- 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
Classifications
-
- 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
- G05D1/0088—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours
Definitions
- the present invention relates to the general field of unmanned aerial vehicles. It belongs to the particular field of autonomous flight operations of aircraft without pilots, especially in a situation of no connection or reduced connection to a control station, with autonomous navigation and behavior embedded in the aircraft.
- NORTHROP GRUMMAN CORPORATION addresses the particular problem of managing the loss of connectivity, or the encounter of unpredictable situations, by their patent US2006106506.
- Their invention includes an evaluation of the performance of the vehicle to declare the unexpected.
- a flight plan to return to the base is provided. This flight plan takes into consideration environmental conditions such as no-fly zones, terrain data, weather forecasts, and alternative control stations to restore communications. It is a question of a limited decisional autonomy able to detect an unforeseen, and to interrupt the mission in the most secure way possible.
- THE BOEING COMPANY details the operation of a system and method of decisional autonomy in its patent EP3101502.
- an apparatus leaves the transmission radius of the station on the ground, it can continue to perform the mission until the communication is restored through a mechanism that can be described as corrective.
- This mechanism performs cycles of verification of the performances of the various components necessary for the functioning of the decisional autonomy.
- a set of corrective actions are performed and a new reassessment cycle is undertaken.
- This system is designed in a logic where the device is normally controlled remotely, and it is only in case of loss of connection that this Corrective system comes into play to maintain a minimum system operation according to the defined objectives, the time that the connection is restored.
- the software components that provide these three main functions identified are: the Behavior Module (103), the Mission Module (104) and the Vision Module (105).
- the Behavior Module (103) is the main module that is started at the beginning of the mission. It uses the Mission Module (104) and the Vision Module (105) to retrieve information and execute commands.
- Module as a software program that can be loaded into a computer's memory and executed by its processors, with a shared static and dynamic data space accessible by all program instructions.
- a Module may consist of submodules sharing the same characteristics and being able to access the data space of the parent module.
- a second secondary computer (102) embarked in the unmanned aircraft can be used in the event of failure of the first computer (100).
- the secondary computer (102) is connected to the main computer (100) by a bidirectional information exchange means such as a connection based on the Internet Protocol standard.
- the secondary computer (102) periodically checks the correct operation of the first computer (100) and takes over in case of detected failure. Only components essential to a safe return to the nearest landing zone are embedded in the secondary computer (102).
- the Behavior Module (103) is implemented in the memory of the computer as a
- each node is a Behavior Tree (200) and each arc between two A to B trees is a call from an Action of Tree A to proceed to the execution of the root of Tree B. Only Tree is the starting point of execution of the Behavior Module (103).
- the number of trees in the graph is at least one; the number of arcs is at least zero.
- the behavior tree model used is based on a tree of nodes representing conditions, actions, and standard instructions that structure the execution of the tree. Each node can take input parameters and return a true / false Boolean value representing the pass / fail states respectively.
- the run progress of the model is based on passes. In each pass, the computer executes the root node and calls the recursive execution of its children.
- the Conditions represent information retrieved from the Mission Module (104), the Vision Module (105), or the Behavior Module (103) that represents internal data, data relayed from the Device Systems. (101) or a combination of both. Conditions can only be true or false.
- Actions represent commands sent to the Mission Module (104), the Vision Module (105), or the Behavior Module (103) that have an internal impact on these modules, an impact on the Systems of the apparatus (101) or a combination of both. Actions are considered true if they succeed and false if they fail.
- the Standard Instructions consist of multi-child scheduling instructions, one-child-only change instructions, and child-only sheet instructions.
- the scheduling instructions supporting several children are: sequential execution of the child nodes and end with failure as soon as a child fails (201); sequential execution of child nodes and end with failure only if all children fail (202); parallel execution of the child nodes (203); random execution of the child nodes and ending with failure as soon as a child fails (204); random execution of child nodes and end with failure only if all children fail (205).
- modification instructions supporting a single child are: Execution switch of the child node (210) at the request of an interrupt instruction (221); Inverter of the execution result of the child node (21 1); Limiter of the number of executions of the child (212); Repeater of the execution of the child node indefinitely (213); Repeater of the execution of the child node until it fails (214); Instruction that forces the success of the child node (215).
- leaf instructions that do not support children are: Wait for a defined period of time (220); Execute an interruption (221); Return true (222); Return False (223); Change data in the global context (224); Execute a tree (thus creating an arc in the graph) (225); Perform an Action (226); Check a Condition (227); Execute a specific reinforcement learning algorithm (107); Check an inference on a neural network (108).
- the Mission Module (104) consists of the following modules: Terrain Model (109), Mission Processor (1 10), Obstacle Detector (1 1 1), Flight Planner (1 12), Scanner Surfaces (1 13), Take-off & Landing (1 14), Object Tracker (115), Special Operations (116), Energy & Time Estimator (117), Advanced Adjustments (1 18), and Communications Center (1 19). All of these modules share the same data space in the computer.
- the Mission Module (104) is the interface between the Behavior Module (103) and the Device Systems (101). It sends discrete commands to the Device Systems (101) in response to the Actions of the Behavior Module (103) and retrieves and consolidates the information retrieved from these systems (101) to feed them to the Conditions.
- the Field Model (109) is a submodule of the Mission Module (104). It keeps in memory and updates if necessary a model of three-dimensional representation of the terrain including but not limited to the elevation data, meshes, the geo-fence volume from which the device must not go out, the no-fly zones, and weather forecasts.
- the Mission Processor (1 10) is a submodule of the Mission Module (104). It remembers and updates as needed information about the tasks and objectives of the mission assigned to the device, as well as the current status of progress in performing tasks.
- the Obstacle Detector (1 1 1) is a submodule of the Mission Module (104). It processes the sensor data of the device and retrieves useful information about the objects surrounding the device in its environment outside the elevation that is already pre-existing in the Terrain Module (109).
- the Flight Planner (1 12) is a submodule of the Mission Module (104). It mainly receives data from the Terrain Model (109) and the Obstacle Detector (1 1 1) and generates optimized flight plans linking any point A to any point B within the geo-fence. These flight plans are used in more complex calculations or transmitted to the autopilot to be executed by the aircraft, depending on the Action initiated by the Behavior Module (103).
- the Surface Scanner (1 13) is a submodule of the Mission Module (104). It mainly receives data from the Terrain Model (109) and the Obstacle Detector (1 1 1) and generates optimized flight plans designed to cover as effectively as possible geographic areas defined in suitable models.
- Take-off & Landing (1-14) is a submodule of the Mission Module (104). It determines the maneuvers necessary to safely succeed the sensitive take-off and landing phases, translates them into commands, and sends them to the autopilot at the request of the Behavior Module Actions (103).
- the Object follower (1 15) is a submodule of the Mission Module (104). It determines the maneuvers necessary to follow an object according to defined parameters, translates them into commands, and sends them to the autopilot at the request of the Actions of the Behavior Module (103).
- Special Maneuvers (1 16) is a submodule of the Mission Module (104). It has the parametric commands necessary to perform special maneuvers. These maneuvers include but are not limited to hovering silently, zigzagging, or dropping a load. These commands are sent to the autopilot at the request of a Behavior Module Action (103).
- the Energy & Time Estimator (1 17) is a submodule of the Mission Module (104). It mainly receives data from the Flight Planner (1 12), Surface Scanner (1 13) and Takeoff & Landing (1 14) modules as well as the onboard sensor history to calculate the energy or time required to complete a task by the Mission Processor (1 10) with an estimate of the error.
- Advanced Adjustments (1 18) is a submodule of the Mission Module (104). It executes in the background a continuous trial & error algorithm that regularly injects changes to the parameters of the autopilot and analyzes the impact on the endurance of the aircraft in order to maximize it.
- the Communication Center (1 19) is a submodule of the Mission Module (104). It records and processes incoming and outgoing messages. Processing consists of encrypting and sending outgoing messages, or decrypting and forwarding incoming messages.
- the Vision Module (105) is composed of a set of submodules that are generally sequentially executed and which are: Image Preprocessor (118), Region Detection (119), and Classification (120).
- the Image Preprocessor (1 18) is a sub-module of the Vision Module (105). It retrieves the last image of the sensor, applies some filters to it and cuts it according to the altitude of the device, its attitude, and the specifications of the sensor. Each cut image is inputted to the Region Detection Module (1 19).
- the Regions Detection (1 19) is a submodule of the Vision Module (105). It performs neural network inference on the input image and returns the detected objects with their positions on the image as a rectangle that are translated into positions on the real world. Each rectangle is then cut off from the image and injected into the Classification Module (120).
- the Classification (120) is a submodule of the Vision Module (105). It runs a set of subroutines that generate information about the object with calculated error rates. These subprograms include but are not limited to neuron network classification inference, Descriptor detection and comparison with an existing Descriptor database, execution of special filters, and recovery of descriptors. information extracted from the filtered image.
- a typical operating method of this system resulting from a specific implementation of the Graph in the Behavior Module with a first Behavior Tree (200) that includes a sequential execution of: Action to arm the autopilot (230); followed by verification of the weather condition acceptable for the flight (250); followed by verification of the Condition of availability of sufficient energy (251), at the Energy & Time Estimator (1 17), to execute all the tasks supplied in the Mission Processor (1 10); tracking, according to the result of the condition, the interruption of the mission or the parallel execution (203) of the processes that represent trees of behavior of: navigation; Communication ; and detection.
- the navigation tree includes a sequential execution (201) of: the Takeoff Action (231) at the Take-off & Landing Module (1 14); followed by a loop execution (214) of the sequence (201): checking the condition that no task remains (252); monitoring the Condition check that the energy is sufficient (251), at the Energy & Time Estimator (1 17), to perform the next task; followed, according to the result of Condition (251), to the interruption of the mission or Action to perform the next task (232).
- the loop (214) terminates upon verification of the Condition that there are no tasks left (252) at the Mission Planner (1 10); Following this, the Landing Action (233) is executed at the Takeoff & Landing Module (1 14).
- the communication tree comprises a loop (214) which executes in sequence (201): Checking the condition for receiving new commands (253) at the communication center (1 19); tracking the execution of the Order Action (234); monitoring the execution of the Mission Resume Action (235) by resuming the last task indicated by the Mission Processor (1 10).
- the detection tree comprises a loop (214) which executes in sequence (201): Launching the detection action (236); followed by the verification of the Condition of the threshold of certainty of the detection of the object (254) required by the mission; followed by, in the case where the result of the Condition (254) is positive, the execution of the Send Alert Action (237) at the Communication Center (1 19).
- the Action to perform the following task is usually a command to the Flight Planner (1 12) or the Surface Scanner (1 13) to calculate a flight plan and execute it at the autopilot level .
- the landing action (233) includes the search for the closest landing zone with the possibility of constraining it to present good conditions (quality of the runway and wind direction, among others) according to the parameters of the landing zone. Entry of the Action.
- the operating method can obviously change depending on the configuration of the behavior graph.
- Other implementations may include and use as an example: an anomaly detection and correction tree, an advanced investigation tree of a detected object, or a navigation tree to the broadband connection point the closest to send large data in case of a critical situation.
- Figure 1 shows a logical representation of the system components object of this invention.
- Figure 2 shows a typical operation of this system, resulting from a specific implementation of the Graph in the Behavior Module (103).
Landscapes
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Radar, Positioning & Navigation (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Game Theory and Decision Science (AREA)
- Business, Economics & Management (AREA)
- Aviation & Aerospace Engineering (AREA)
- Health & Medical Sciences (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Traffic Control Systems (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| MAMA41572 | 2017-12-05 | ||
| MA41572A MA41572B1 (fr) | 2017-12-05 | 2017-12-05 | Navigation et comportement autonomes de vehicule sans pilote sans liaison avec la station de controle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2019112407A1 true WO2019112407A1 (fr) | 2019-06-13 |
Family
ID=65019551
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/MA2018/050013 Ceased WO2019112407A1 (fr) | 2017-12-05 | 2018-12-04 | Navigation et comportement autonomes de véhicule sans pilote sans liaison avec la station de contrôle |
Country Status (2)
| Country | Link |
|---|---|
| MA (1) | MA41572B1 (fr) |
| WO (1) | WO2019112407A1 (fr) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111353606A (zh) * | 2020-02-29 | 2020-06-30 | 中国电子科技集团公司第五十二研究所 | 一种基于模糊决策树的深度强化学习空战博弈解释方法和系统 |
| CN111914412A (zh) * | 2020-07-21 | 2020-11-10 | 同济大学 | 一种基于错误注入器的自动驾驶性能局限测试系统及方法 |
| CN112099525A (zh) * | 2020-08-31 | 2020-12-18 | 北京航空航天大学 | 一种航天器编队飞行低通讯连通保持协同控制方法 |
| CN113110114A (zh) * | 2021-05-24 | 2021-07-13 | 北京润科通用技术有限公司 | 一种超实时联合仿真的调度方法及装置 |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060106506A1 (en) | 2004-11-16 | 2006-05-18 | Nichols William M | Automatic contingency generator |
| US20110184604A1 (en) | 2005-02-16 | 2011-07-28 | Lockheed Martin Corporation | Hierarchical contingency management system for mission planners |
| US20150051783A1 (en) * | 2012-03-22 | 2015-02-19 | Israel Aerospace Industries Ltd. | Planning and monitoring of autonomous-mission |
| WO2016033796A1 (fr) * | 2014-09-05 | 2016-03-10 | SZ DJI Technology Co., Ltd. | Sélection de mode de vol basée sur le contexte |
| US20160187882A1 (en) * | 2011-08-16 | 2016-06-30 | Unmanned Innovation, Inc. | Modular flight management system incorporating an autopilot |
| EP3101502A2 (fr) | 2015-06-05 | 2016-12-07 | The Boeing Company | Prise de décision de véhicule aérien télépiloté autonome |
-
2017
- 2017-12-05 MA MA41572A patent/MA41572B1/fr unknown
-
2018
- 2018-12-04 WO PCT/MA2018/050013 patent/WO2019112407A1/fr not_active Ceased
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060106506A1 (en) | 2004-11-16 | 2006-05-18 | Nichols William M | Automatic contingency generator |
| US20110184604A1 (en) | 2005-02-16 | 2011-07-28 | Lockheed Martin Corporation | Hierarchical contingency management system for mission planners |
| US20160187882A1 (en) * | 2011-08-16 | 2016-06-30 | Unmanned Innovation, Inc. | Modular flight management system incorporating an autopilot |
| US20150051783A1 (en) * | 2012-03-22 | 2015-02-19 | Israel Aerospace Industries Ltd. | Planning and monitoring of autonomous-mission |
| WO2016033796A1 (fr) * | 2014-09-05 | 2016-03-10 | SZ DJI Technology Co., Ltd. | Sélection de mode de vol basée sur le contexte |
| EP3101502A2 (fr) | 2015-06-05 | 2016-12-07 | The Boeing Company | Prise de décision de véhicule aérien télépiloté autonome |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111353606A (zh) * | 2020-02-29 | 2020-06-30 | 中国电子科技集团公司第五十二研究所 | 一种基于模糊决策树的深度强化学习空战博弈解释方法和系统 |
| CN111353606B (zh) * | 2020-02-29 | 2022-05-03 | 中国电子科技集团公司第五十二研究所 | 一种基于模糊决策树的深度强化学习空战博弈方法和系统 |
| CN111914412A (zh) * | 2020-07-21 | 2020-11-10 | 同济大学 | 一种基于错误注入器的自动驾驶性能局限测试系统及方法 |
| CN111914412B (zh) * | 2020-07-21 | 2023-07-04 | 同济大学 | 一种基于错误注入器的自动驾驶性能局限测试系统及方法 |
| CN112099525A (zh) * | 2020-08-31 | 2020-12-18 | 北京航空航天大学 | 一种航天器编队飞行低通讯连通保持协同控制方法 |
| CN113110114A (zh) * | 2021-05-24 | 2021-07-13 | 北京润科通用技术有限公司 | 一种超实时联合仿真的调度方法及装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| MA41572B1 (fr) | 2019-08-30 |
| MA41572A1 (fr) | 2019-06-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US12405615B2 (en) | Cloud and hybrid-cloud flight vehicle and robotic control system AI and ML enabled cloud-based software and data system method for the optimization and distribution of flight control and robotic system solutions and capabilities | |
| WO2019112407A1 (fr) | Navigation et comportement autonomes de véhicule sans pilote sans liaison avec la station de contrôle | |
| US8265818B2 (en) | Open architecture command system | |
| EP3489929B1 (fr) | Système électronique de télé-pilotage de drones, procédé et programme d'ordinateur associés | |
| US10921812B2 (en) | System and method for rendezvous coordination of an autonomous automotive vehicle with an ambulance | |
| CN110723150A (zh) | 处理自主车辆处的乘坐者服务 | |
| FR3010809A1 (fr) | Procede et dispositif de gestion automatique a bord d'un aeronef de messages audio de controle aerien. | |
| US12260616B1 (en) | Multi-task machine learning model for event detection | |
| WO2021105055A1 (fr) | Dispositif et procede d'aide a la decision pour la gestion de conflits aeriens | |
| WO2020127703A1 (fr) | Dispositif de pilotage destine a etre integre dans un aeronef preexistant | |
| US20240231359A1 (en) | Advanced flight processing system and/or method | |
| CA3020086A1 (fr) | Procede de controle de la restitution d'alerte(s) et/ou de procedure(s) de reconfiguration systeme(s), produit programme d'ordinateur et systeme de controle associes | |
| WO2023137285A1 (fr) | Systèmes et procédés d'optimisation de trajet de vol en nuage pour véhicules aériens sans pilote | |
| US11392139B2 (en) | Method, apparatus and control system for controlling mobile robot | |
| Schirmer et al. | Considerations of artificial intelligence safety engineering for unmanned aircraft | |
| EP3232417A1 (fr) | Securisation du sequencement du plan de vol d'un aeronef | |
| WO2021094178A1 (fr) | Système de pilotage de drone et procédé de pilotage associé | |
| WO2023135202A1 (fr) | Système de sélection de trajectoire d'un aéronef muni d'un système de navigation et guidage automatique sur une trajectoire | |
| Shishkov et al. | Using drones for resilience: a system of systems perspective | |
| EP4078559A1 (fr) | Gestion de l'encombrement spatial autour de la trajectoire d'un vehicule | |
| US20230225036A1 (en) | Power conservation tools and techniques for emergency vehicle lighting systems | |
| US11988749B2 (en) | System and method for hybrid LiDAR segmentation with outlier detection | |
| FR3090979A1 (fr) | Système de pilotage alternatif destiné à être intégré dans un aéronef préexistant | |
| Smirnov et al. | Knowledge elicitation and representation for module based perceptual capabilities onboard UAVs | |
| FR3092428A1 (fr) | Dispositif d’acquisition et d’analyse destiné à être intégré dans un aéronef préexistant |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 18833729 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: 18833729 Country of ref document: EP Kind code of ref document: A1 |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 15/11/2021) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 18833729 Country of ref document: EP Kind code of ref document: A1 |