WO2024174437A1 - Adaptive-parameter local path planning method and system for mobile robot - Google Patents

Abstract

An adaptive-parameter local path planning method and system for a mobile robot. When predicted trajectories are evaluated by means of a trajectory evaluation model and a path is planned, all predicted trajectories corresponding to a plurality of velocity sets are traversed to determine the number of predicted trajectories in which an obstacle is present among all the predicted trajectories; and if the number of predicted trajectories in which an obstacle is present is greater than a set threshold value, the proportions of an orientation cost and a velocity difference cost in the trajectory evaluation model are reduced, and a sampling velocity is calculated on the basis of the updated proportions. By means of adaptively adjusting the proportions of the orientation cost and velocity difference cost in the trajectory evaluation model, the problem of a robot being trapped in a local optimal solution is effectively avoided without increasing computational burden.

Description

A method and system for local path planning of mobile robots with adaptive parameters

The present invention claims the priority of the Chinese patent application filed with the Chinese Patent Office on February 21, 2023, with application number 202310138722.2 and invention name “A local path planning method and system for mobile robots with adaptive parameters”, the entire contents of which are incorporated into the present invention by reference.

Technical Field

The present invention belongs to the technical field of path optimization, and in particular relates to a method and system for local path planning of a mobile robot with adaptive parameters.

Background Art

Path planning algorithms are mainly divided into global path planning and local path planning. Among them, the Dynamic Window Approach (DWA) is a typical local path planner. The Dynamic Window Approach is more inclined to choose a trajectory moving towards the target direction. The Dynamic Window Approach samples the linear velocity and angular velocity in the velocity space, and predicts the trajectory of the next time interval based on the robot's kinematic model. It scores the trajectory to be evaluated, thereby obtaining a safer and smoother optimal local path.

The inventors found that when using the traditional robot local path planning algorithm, when the robot is in a position where there is an obstacle in front of it, the desired trajectory not only needs to move towards the target, but also needs to move away from the obstacle. This contradiction causes the robot to fall into a local minimum point and the problem of being trapped in a local optimal solution. Specifically, it constantly cycles back and forth or even stagnates near this position, resulting in poor obstacle avoidance effect.

Summary of the invention

In order to solve the above problems, the present invention proposes a method and system for local path planning of a mobile robot with adaptive parameters. The present invention utilizes a robot local path planning algorithm with adaptive parameters to avoid the robot falling into local optimum when performing local path planning.

In order to achieve the above object, the present invention is implemented through the following technical solutions:

In a first aspect, the present invention provides a method for local path planning of a mobile robot with adaptive parameters, comprising:

Sampling in the velocity space to obtain multiple groups of velocity sets;

Perform trajectory prediction for each set of speeds to obtain a predicted trajectory;

The predicted trajectory is evaluated by a preset trajectory evaluation model, and the path is planned by adjusting the parameters in the trajectory evaluation model to obtain the optimal path;

The trajectory evaluation model includes a direction cost, a speed difference cost and a total trajectory cost; all predicted trajectories corresponding to multiple groups of speed sets are traversed to determine the number of obstacles encountered in all predicted trajectories. If the number of predicted trajectories encountering obstacles is greater than a set threshold, the score proportion of the direction cost and the score proportion of the speed difference cost in the trajectory evaluation model are reduced, and the sampling speed is calculated based on the updated proportions.

Furthermore, the orientation cost is the angle between the current position of the robot and the local target position, and the current target orientation of the robot; the speed difference cost is the difference between the current speed of the robot and the maximum speed of the robot; the total trajectory cost is the sum of the costs corresponding to all trajectory points in the cost map in a set of sampled speeds.

Furthermore, the trajectory evaluation model is equal to the product of the adaptive coefficient of the heading cost, the heading cost, and the score proportion corresponding to the heading cost plus the adaptive coefficient of the speed difference cost. The sum of the products of the adaptation coefficient, the speed difference cost, and the score proportion corresponding to the speed difference cost, plus the sum of the products of the total generation value of the trajectory and the score proportion corresponding to the total generation value of the trajectory.

Furthermore, the adaptive coefficient of the heading cost and the adaptive coefficient of the speed difference cost are both the largest value between 0.1 and the predicted value; the predicted value is equal to the ratio of the total number of predicted trajectories and the number of trajectories encountering obstacles to the total number of predicted trajectories.

Furthermore, it is determined whether a collision occurs when the first control cycle trajectory point in the predicted trajectory is emergency stopped. If so, the corresponding speed sampling is discarded. Otherwise, the corresponding cost of the trajectory point in each control cycle in the cost map is accumulated.

Furthermore, if the number of predicted trajectories encountering obstacles is less than or equal to the set threshold, the current point of the robot is connected to the local target point; if there is obstacle information on the line connecting the current point of the robot and the local target point, the global planner is used to replan the path, and the planned path is put into the local planning target point set, the local target point is updated, substituted into the trajectory evaluation model, and the optimal sampling speed is calculated.

Furthermore, the global planner is used to search for path points to the target point;

Use Floyd's path smoothing algorithm to set the key points of local path planning;

Select the target point among the key points;

When the machine is moving, it builds a local map and trajectory evaluation model, and adjusts the local planning scheme according to whether there are obstacles between the current point and the target point.

In a second aspect, the present invention further provides a mobile robot local path planning system with adaptive parameters, comprising:

The sampling module is configured to: sample in the velocity space to obtain multiple velocity sets;

The prediction module is configured to: perform trajectory prediction on each set of speed sets to obtain a predicted trajectory;

The planning module is configured to: evaluate the predicted trajectory through a preset trajectory evaluation model, and plan the path by adjusting parameters in the trajectory evaluation model to obtain an optimal path;

The trajectory evaluation model includes a direction cost, a speed difference cost and a total trajectory cost; all predicted trajectories corresponding to multiple groups of speed sets are traversed to determine the number of obstacles encountered in all predicted trajectories. If the number of predicted trajectories encountering obstacles is greater than a set threshold, the score proportion of the direction cost and the score proportion of the speed difference cost in the trajectory evaluation model are reduced, and the sampling speed is calculated based on the updated proportions.

In a third aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for local path planning of a mobile robot with adaptive parameters described in the first aspect.

In a fourth aspect, the present invention further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the method for local path planning of a mobile robot with adaptive parameters described in the first aspect are implemented.

Compared with the prior art, the present invention has the following beneficial effects:

In the present invention, when evaluating the predicted trajectory and planning the path through the trajectory evaluation model, all predicted trajectories corresponding to multiple groups of speed sets are traversed to determine the number of obstacles encountered in all predicted trajectories. If the number of predicted trajectories that encounter obstacles is greater than a set threshold, the proportion of the orientation cost and the proportion of the speed difference cost in the trajectory evaluation model are reduced. The sampling speed is calculated based on the updated proportion, and the proportion of orientation cost and speed difference cost in the trajectory evaluation model is adaptively adjusted to effectively avoid the robot from falling into the local optimal solution without increasing the computational burden.

BRIEF DESCRIPTION OF THE DRAWINGS

The drawings in the specification that constitute a part of this embodiment are used to provide a further understanding of this embodiment. The schematic embodiments of this embodiment and their descriptions are used to explain this embodiment and do not constitute improper limitations on this embodiment.

FIG1 is a flow chart of Embodiment 1 of the present invention;

FIG2 is a local planner solution of Embodiment 1 of the present invention.

DETAILED DESCRIPTION

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

Embodiment 1:

This embodiment provides a method for local path planning of a mobile robot with adaptive parameters, including:

S1. Use the global planner to search for path points to the target point;

S2. Using the Floyd path smoothing algorithm, the path points on a straight line from the starting point and exceeding a certain threshold are set as local path planning key points. It can be understood that there are no obstacles on this straight line; the straight line refers to the line connecting the current point and the local target point, and the threshold can be understood as the shortest distance between the local path key points;

S3, select the local target point to be tracked according to the current position of the robot and the threshold of the target point to be tracked, and select the target point from the key points; specifically, the selection process needs to meet the following conditions: the distance from the current point exceeds a certain threshold, and the point is selected as the local target point for the first time;

S4, when the robot is walking, by building a local map and sampling evaluation model, the local planning scheme is adjusted according to whether there are obstacles between the current point and the local target point;

Specifically, when building a local map, in this embodiment, it is selected to first delete the data information of the previous frame, and then construct the obstacle information according to the data of the current frame; the sampling evaluation model is a trajectory evaluation model.

S5, sending robot walking information;

It can be understood that the robot is controlled by the adjusted local planning scheme.

Through this embodiment, different obstacle stopping and obstacle bypassing distances can be performed according to different needs of different mobile robots, and the machine can be effectively prevented from falling into a local optimal solution without increasing the computational burden.

As described in the background technology, when the traditional dynamic window method is used for local path planning, when there is an obstacle in front of the robot, the desired trajectory needs to move not only toward the target but also away from the obstacle. This contradiction causes the robot to fall into a local minimum point. In order to solve the above problem, in the local path planning method of the mobile robot with adaptive parameters, the following improvements are made:

Sampling in the velocity space to obtain multiple groups of velocity sets;

Perform trajectory prediction on each set of speed sets to obtain a predicted trajectory. Trajectory prediction can be achieved through conventional techniques, which will not be described in detail here.

The predicted trajectory is evaluated by the preset trajectory evaluation model, and the trajectory is adjusted Evaluate the parameters in the model, plan the path, and obtain the optimal path;

The trajectory evaluation model includes a direction cost, a speed difference cost and a total trajectory cost; all predicted trajectories corresponding to multiple groups of speed sets are traversed to determine the number of obstacles encountered in all predicted trajectories. If the number of predicted trajectories encountering obstacles is greater than a set threshold, the score proportion of the direction cost and the score proportion of the speed difference cost in the trajectory evaluation model are reduced, and the sampling speed is calculated based on the updated proportions.

When evaluating the predicted trajectory and planning the path through the trajectory evaluation model, all predicted trajectories corresponding to multiple groups of speed sets are traversed to determine the number of obstacles encountered in all predicted trajectories. If the number of predicted trajectories encountering obstacles is greater than the set threshold, the proportion of orientation cost and the proportion of speed difference cost in the trajectory evaluation model are reduced, and the sampling speed is calculated based on the updated proportions. By adaptively adjusting the proportion of orientation cost and the proportion of speed difference cost in the trajectory evaluation model, the problem of the robot falling into a local minimum point is effectively avoided without increasing the computational burden.

The orientation cost A is the angle between the current position of the robot and the local target position, as well as the current target orientation of the robot, indicating the planned direction of the robot; the score proportion As corresponding to the orientation cost has a default minimum value of 0.1, and the default value of the adaptive coefficient Aa of the orientation cost is 1. The speed difference cost B is the difference between the current speed v of the robot and the maximum speed Mv of the robot, indicating that the faster the robot is, the closer it is to the target. The default minimum value of the score proportion Bs corresponding to the speed difference cost is 0.1, and the default value of the adaptive coefficient Ba of the speed difference cost is 1. The total trajectory cost C is the sum of the costs corresponding to all trajectory points in a set of sampled speeds in the cost map, and the score proportion Cs corresponding to the total trajectory cost. Among them, the cost map refers to the obstacle information in the static map combined with the machine The robot’s configuration space,simplifies the robot into a map that can be represented as a point in the cost map, which can be understood as an obstacle point, and the obstacle is inflated according to the robot shape.

Among them, the total cost value of the trajectory C = 1-adaptive coefficient of the orientation cost Aa*score proportion As corresponding to the orientation cost-adaptive coefficient Ba of the speed difference cost*score proportion Bs corresponding to the speed difference cost. Score proportion As corresponding to the orientation cost + score proportion Bs corresponding to the speed difference cost + score proportion Cs corresponding to the total cost value of the trajectory = 1; * represents a multiplication sign.

The trajectory evaluation model is equal to the product of the adaptive coefficient of the orientation cost, the orientation cost and the score ratio corresponding to the orientation cost plus the sum of the products of the adaptive coefficient of the speed difference cost, the speed difference cost and the score ratio corresponding to the speed difference cost, plus the sum of the products of the total generation value of the trajectory and the score ratio corresponding to the total generation value of the trajectory; specifically, the final trajectory score = adaptive coefficient of orientation cost Aa*orientation cost A*score ratio corresponding to orientation cost As+adaptive coefficient of speed difference cost Ba*speed difference cost B*score ratio corresponding to speed difference cost Bs+total generation value of trajectory C*score ratio corresponding to the total generation value of trajectory Cs; * represents the multiplication sign.

The speed sampling space is Vs∩Vd, and trajectory prediction is performed for each set of speeds in the sampling space;
Vs={(sv, sw)|sv∈[Nv, Mv]Λsw∈[Nw, Mw]}
Vd={(sv, sw)|sv∈[Cv-Av*St, Cv+Av*St]Λsw∈[Cw-
Aw*St，Cw+Aw*St]}

Among them, Vs is the speed that the robot can reach under the motion model; sv is the sampling speed; sw is the sampling angular velocity; Nv is the minimum speed; Mv is the maximum speed; sw is the sampling angular velocity; Nw is the minimum angular velocity; Mw is the maximum angular velocity; Vd is the robot dynamics model The sampling speed that the machine can achieve; Cv is the current speed; Av is the linear acceleration; St is the sampling time, and the sampling time St is greater than the control time Ct; Cw is the current angular velocity; Aw is the angular acceleration.

For the first control cycle trajectory point in the predicted trajectory, an emergency stop is performed to determine whether a collision will occur. If a collision occurs, the group of velocity samples is discarded. Otherwise, the corresponding cost of the trajectory point in each control cycle in the cost map is accumulated to record whether the trajectory touches an obstacle.

The total number of trajectories of all group speeds is SUM, and the set of trajectories that encounter obstacles is determined. The number of trajectories that encounter obstacles is N. If the set size exceeds the threshold, it is determined that the machine will enter a dense obstacle area, and the speed difference and orientation ratio are reduced; trajectory evaluation model, normalized trajectory score, update the corresponding score ratio, calculate the optimal sampling speed, it can be understood that each sampled speed and angular velocity can get a trajectory information, each trajectory information can get a score, and the sampling speed and angular velocity corresponding to the trajectory with the smallest score is the optimal speed; the adaptive coefficient of the orientation cost and the adaptive coefficient of the speed difference cost are both the largest value between 0.1 and the predicted value; the predicted value is equal to the total number of predicted trajectories and the number of trajectories encountering obstacles, and then the ratio of the total number of predicted trajectories; specifically, the adaptive coefficient of the orientation cost Aa = max((SUM-N)/SUM, 0.1); the adaptive coefficient of the speed difference cost Ba = max((SUM-N)/SUM, 0.1).

If the set size does not exceed the threshold, connect the current point and the local target point. If there is obstacle information on the connection line, use the global planner to replan the path and limit the path search time of the global planner to avoid the path search taking too long. Put the path into the local planning target point set, update the local target point, substitute it into the trajectory evaluation model, and calculate the optimal sampling speed.

The optimal sampling speed is sent to the chassis of the mobile robot to achieve robot control and path planning.

Embodiment 2:

This embodiment provides a local path planning system for a mobile robot with adaptive parameters, including:

The sampling module is configured to: sample in the velocity space to obtain multiple velocity sets;

The prediction module is configured to: perform trajectory prediction on each set of speed sets to obtain a predicted trajectory;

The planning module is configured to: evaluate the predicted trajectory through a preset trajectory evaluation model, and plan the path by adjusting parameters in the trajectory evaluation model to obtain an optimal path;

The trajectory evaluation model includes a direction cost, a speed difference cost and a total trajectory cost; all predicted trajectories corresponding to multiple groups of speed sets are traversed to determine the number of obstacles encountered in all predicted trajectories. If the number of predicted trajectories encountering obstacles is greater than a set threshold, the score proportion of the direction cost and the score proportion of the speed difference cost in the trajectory evaluation model are reduced, and the sampling speed is calculated based on the updated proportions.

The working method of the system is the same as the local path planning method of the mobile robot with adaptive parameters in Example 1, and will not be repeated here.

Embodiment 3:

This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the method for local path planning of a mobile robot with adaptive parameters described in Example 1 are implemented.

Embodiment 4:

This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the method for local path planning of a mobile robot with adaptive parameters described in Example 1 are implemented.

The above description is only a preferred embodiment of the present embodiment and is not intended to limit the present embodiment. For those skilled in the art, the present embodiment may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment shall be included in the protection scope of the present embodiment.

Claims (10)

A method for local path planning of a mobile robot with adaptive parameters, characterized by comprising: Sampling in the velocity space to obtain multiple groups of velocity sets; Perform trajectory prediction for each set of speeds to obtain a predicted trajectory; The predicted trajectory is evaluated by a preset trajectory evaluation model, and the path is planned by adjusting the parameters in the trajectory evaluation model to obtain the optimal path; The trajectory evaluation model includes orientation cost, speed difference cost and total trajectory cost; all predicted trajectories corresponding to multiple groups of speed sets are traversed to determine the number of obstacles encountered in all predicted trajectories. If the number of predicted trajectories encountering obstacles is greater than a set threshold, the score proportion of the orientation cost and the score proportion of the speed difference cost in the trajectory evaluation model are reduced, and the sampling speed is calculated based on the updated proportions. A method for local path planning of a mobile robot with adaptive parameters as described in claim 1, characterized in that the orientation cost is the angle between the current position of the robot and the local target position, and the current target orientation of the robot; the speed difference cost is the difference between the current speed of the robot and the maximum speed of the robot; the total trajectory cost is the sum of the corresponding costs of all trajectory points in the cost map in a set of sampled speeds. A local path planning method for a mobile robot with adaptive parameters as described in claim 1, characterized in that the trajectory evaluation model is equal to the product of the adaptive coefficient of the orientation cost, the orientation cost, and the score proportion corresponding to the orientation cost, plus the sum of the products of the adaptive coefficient of the speed difference cost, the speed difference cost, and the score proportion corresponding to the speed difference cost, plus the sum of the products of the total cost of the trajectory and the score proportion corresponding to the total cost of the trajectory. A method for local path planning of a mobile robot with adaptive parameters as described in claim 3, characterized in that the adaptive coefficient of the orientation cost and the adaptive coefficient of the speed difference cost are both the largest value between 0.1 and the predicted value; the predicted value is equal to the ratio of the total number of predicted trajectories and the number of trajectories encountering obstacles to the total number of predicted trajectories. A method for local path planning of a mobile robot with adaptive parameters as described in claim 1, characterized in that it is determined whether a collision occurs when an emergency stop is performed on the first control cycle trajectory point in the predicted trajectory. If so, the corresponding speed sampling is discarded, otherwise, the corresponding cost of the trajectory point in each control cycle in the cost map is accumulated. A local path planning method for a mobile robot with adaptive parameters as described in claim 1, characterized in that if the number of predicted trajectories encountering obstacles is less than or equal to a set threshold, the current point of the robot is connected to the local target point; if there is obstacle information on the line connecting the current point of the robot and the local target point, the path is replanned using a global planner, and the planned path is placed in a local planning target point set, the local target point is updated, substituted into the trajectory evaluation model, and the optimal sampling speed is calculated. A method for local path planning of a mobile robot with adaptive parameters as claimed in claim 6, characterized in that a global planner is used to search for path points to a target point; Use Floyd's path smoothing algorithm to set the key points of local path planning; Select the target point among the key points; When the machine is moving, it builds a local map and trajectory evaluation model, and adjusts the local planning scheme according to whether there are obstacles between the current point and the target point. A local path planning system for a mobile robot with adaptive parameters, characterized by comprising: The sampling module is configured to: sample in the velocity space to obtain multiple velocity sets; The prediction module is configured to: perform trajectory prediction on each set of speeds to obtain a predicted trajectory; The planning module is configured to: evaluate the predicted trajectory through a preset trajectory evaluation model, and plan the path by adjusting parameters in the trajectory evaluation model to obtain an optimal path; The trajectory evaluation model includes a direction cost, a speed difference cost and a total trajectory cost; all predicted trajectories corresponding to multiple groups of speed sets are traversed to determine the number of obstacles encountered in all predicted trajectories. If the number of predicted trajectories encountering obstacles is greater than a set threshold, the score proportion of the direction cost and the score proportion of the speed difference cost in the trajectory evaluation model are reduced, and the sampling speed is calculated based on the updated proportions. A computer-readable storage medium having a computer program stored thereon, characterized in that when the program is executed by a processor, the steps of the method for local path planning of a mobile robot with adaptive parameters as described in any one of claims 1 to 7 are implemented. An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the method for local path planning of a mobile robot with adaptive parameters as described in any one of claims 1 to 7 are implemented.