A method and system for multi-robot pursuit-evasion obstacle avoidance based on thiesen polygons

Abstract

The application discloses a kind of method and system for multiple robots to avoid obstacles based on thiesen polygon, the method comprises: initialization limited space environment, obtain selected bounded region, selected bounded region is divided using thiesen polygon, distribute pursuit target and pursuit target point, using Delaunay triangle to region segmentation corrosion, obtain safe movement area, based on pursuit target point and safe movement area obtain pursuit obstacle avoidance direction control pointing point.The system includes: initialization module, thiesen polygon module, task allocation module, safe area division module, pursuit direction point module and escape direction point module.Through using the application, a series of problems caused by using reinforcement learning, estimation prediction and other methods can be avoided, and the robot's collision avoidance and pursuit direction selection under the condition of multiple obstacles can be realized with lower algorithm complexity.The application can be widely applied to the field of robot pursuit obstacle avoidance technology.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of robot pursuit and obstacle avoidance, and particularly relates to a method and system for multi-robot pursuit and obstacle avoidance based on Thiessen polygons. BACKGROUND

[0002] A multi-robot system is a kind of mobile distributed system, which has high density characteristics and is robust, scalable and flexible, so it has more advantages than single robots in dealing with large-scale tasks. There are many challenges in applying multi-robot systems to complete multi-target pursuit and obstacle avoidance, which need to solve the problem of robot collision avoidance and obstacle avoidance in multi-obstacle conditions and the problem of multi-robot direction selection in multi-obstacle space. The existing technology has a robot pursuit method based on multi-agent reinforcement learning, which makes the pursuit process more efficient and reliable by constructing a two-pursuit-one-escape environment, constructing a Markov model, obtaining a two-pursuit-one-escape network model, and extending a multi-pursuit-multi-escape strategy. However, this learning algorithm often relies on a large number of learning samples and a long training and debugging process. Moreover, due to the errors between simulation and reality and the problems of communication control, the excellent sample selection in simulation has errors in experiments. There is also a target estimation and prediction algorithm to estimate the position of the escaping robot in real time, obtain the probability density function of the position distribution of the escaping robot, and constantly move to the estimated position of the escaping robot to finally achieve capture. However, in the motion prediction control guided by the machine learning algorithm, there are limitations inherent to model predictive control, which restricts the optimization of the motion scheme and easily falls into a local optimal state, which cannot cooperate with other robots to achieve global optimization. SUMMARY

[0003] To solve the above technical problems, the purpose of the present application is to provide a method and system for multi-robot pursuit and obstacle avoidance based on Thiessen polygons, which can avoid a series of problems caused by the use of reinforcement learning, estimation and prediction methods, and achieve robot collision avoidance and obstacle avoidance and pursuit direction selection under multi-obstacle conditions with lower algorithm complexity.

[0004] The first technical solution adopted by the present application is: a method for multi-robot pursuit and obstacle avoidance based on Thiessen polygons, comprising the following steps:

[0005] Initializing the limited space environment to obtain a selected bounded area;

[0006] Dividing the pursuit and escape robots based on the selected bounded area into Thiessen polygons to obtain Thiessen polygon partitions;

[0007] Assigning targets to the Thiessen polygon partitions to obtain the pursuit targets and pursuit target points of the pursuit robots;

[0008] Segmenting the selected bounded area, the selected pursuit-escape robot and the obstacles to obtain a safety area;

[0009] Combining the pursuit target point and the safety area to obtain a direction control pointing point for the pursuit robot to avoid obstacles;

[0010] Combining the pursuit robot position and the safety area to obtain a direction control pointing point for the escape robot to avoid obstacles.

[0011] Further, the step of initializing the limited space environment to obtain the selected bounded area specifically includes:

[0012] Setting a rectangular space containing several obstacles and pursuit-escape robots to obtain a rectangular bounded area;

[0013] Obtaining the positions of the obstacles, the pursuit robot and the escape robot in the rectangular bounded area to obtain a bounded area layout;

[0014] Setting a safety distance of the obstacles and a capture condition of the escape robot for the bounded area layout to obtain the selected bounded area.

[0015] Through the preferred step, the robot activity range is limited in the bounded area, which reduces the algorithm complexity and the time required for successful pursuit.

[0016] Further, the step of performing a Voronoi polygon division on the pursuit-escape robot based on the selected bounded area specifically includes:

[0017] Connecting the pursuit-escape robots as target points to obtain a Delaunay triangulation graph;

[0018] Calculating the circumcircle centers of each Delaunay triangle and connecting adjacent circle centers to obtain a Voronoi polygon of the triangulation edge;

[0019] Making a vertical bisector of the Voronoi polygon of the triangulation edge intersect the selected bounded area to obtain a Voronoi polygon division.

[0020] Through the preferred step, an intuitive connection is established between the discrete pursuit-escape robot points in the selected bounded area, which facilitates the allocation of the best pursuit target.

[0021] Furthermore, in the step of assigning targets to the Thiessen polygon partitions to obtain the pursuit targets and target points for the pursuit robot, the pursuit target is the escape robot within the Thiessen polygon partition that has the smallest distance between the escape robot and the pursuit robot. When there are adjacent edges in the Thiessen polygon, the midpoint of the adjacent edge of the Thiessen polygon formed by the pursuit robot and the pursuit target is taken as the pursuit target point; when there are no adjacent edges in the Thiessen polygon, the position point of the pursuit target is taken as the pursuit target point. The formulas for the pursuit targets and target points of the pursuit robot are expressed as follows:

[0022]

[0023] Where RidgeEdge(i,j) are the adjacent edges of the Thiessen polygon. In order to capture the target, N represents the distance between the escaping robot and the pursuing robot. P In order to hunt down the robot swarm, N E A group of escaped robots.

[0024]

[0025] in To track down the target point, MidPoint is the midpoint. The adjacent edges of the Thiessen polygon formed by the pursuing robot and the pursuing target. The location of the target being pursued.

[0026] This optimization process determined the optimal match between the pursuit robot and the escape robot in the absence of obstacles.

[0027] Furthermore, the step of dividing the selected bounded area, the selected pursuit robot, and obstacles into a safe area specifically includes:

[0028] Construct the Deloitte triangle based on obstacles and the chasing robot;

[0029] Traverse the adjacent points of the pursuit robot, divide the Deloitte triangle, and obtain the quasi-safe area;

[0030] A safe distance corrosion is performed on the safe area to obtain the safe area.

[0031] This optimization process ensures a safe operating area for the robot, preventing collisions between robots and between the robot and obstacles.

[0032] Furthermore, in the step of combining the target point and the safe area to obtain the directional control pointing point for the pursuit robot to avoid obstacles, the directional control pointing point for the pursuit robot to avoid obstacles is the point within the safe area that is closest to the target point, and its formula is expressed as follows:

[0033]

[0034] Front i (t) represents the target point for pursuit direction control, realV i (t) represents the safe zone. To pursue the target location.

[0035] This optimization process determines the direction of the pursuit robot's obstacle avoidance.

[0036] Furthermore, in the step of combining the target point and the safe area to obtain the directional control pointing point for the escape robot's obstacle avoidance, the directional control pointing point for the escape robot's obstacle avoidance is the point within the safe area that is furthest from the position of the pursuing robot, and its formula is expressed as follows:

[0037]

[0038] Front j (t) represents the escape direction control point, realV i (t) represents the safe zone. To track the robot's location.

[0039] This optimization process determines the direction in which the escape robot avoids obstacles.

[0040] The second technical solution adopted in this invention is: a multi-robot pursuit and obstacle avoidance system based on Thiessen polygons, comprising:

[0041] The initialization module is used to initialize the finite space environment to obtain the selected bounded region;

[0042] The Thiessen polygon module is used to divide the pursuit robot into Thiessen polygon partitions based on a selected bounded region.

[0043] The task allocation module is used to allocate targets according to the Thiessen polygon partitions, so as to obtain the pursuit targets and pursuit target points of the pursuit robot;

[0044] The safe zone division module is used to divide a selected bounded area, a selected pursuit robot, and obstacles to obtain a safe zone.

[0045] The pursuit direction point module is used to combine the pursuit target point and the safe area to obtain the direction control pointing point for the pursuit robot to pursue and avoid obstacles;

[0046] The escape direction point module is used to combine the position of the pursuit robot and the safe zone to obtain the directional control pointing point for the escape robot to escape and avoid obstacles.

[0047] The beneficial effects of the method and system of this invention are as follows: In the case of limited space and a certain number of obstacles, this invention uses Thiessen polygons to divide the region, allocates the target point for pursuit and escape, and uses Delaunay triangles to segment and erode the region to obtain a safe movement region. Based on the target point for pursuit and escape and the safe movement region, the direction control pointing point for pursuit and obstacle avoidance is obtained. This avoids a series of problems caused by using reinforcement learning, estimation and prediction methods, and ensures the safety of the robot's next action. It achieves collision avoidance and obstacle avoidance and pursuit and escape direction selection of the robot under multi-obstacle conditions with lower algorithm complexity. Attached Figure Description

[0048] Figure 1 This is a flowchart of the steps of a multi-robot obstacle avoidance method based on Thiessen polygons according to the present invention;

[0049] Figure 2 This is a structural block diagram of a multi-robot obstacle avoidance system based on Thiessen polygons according to the present invention;

[0050] Figure 3 This is a schematic diagram of the Thiessen polygon partitioning method for multi-robot pursuit and obstacle avoidance based on Thiessen polygons according to the present invention.

[0051] Figure 4 This is a quasi-safe zone map of a multi-robot pursuit and obstacle avoidance method based on Thiessen polygons according to the present invention;

[0052] Figure 5 This is a safety area diagram of a multi-robot pursuit and obstacle avoidance method based on Thiessen polygons according to the present invention. Detailed Implementation

[0053] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only for ease of explanation and do not limit the order of the steps. The execution order of each step in the embodiments can be adapted according to the understanding of those skilled in the art.

[0054] Reference Figure 1 This invention provides a method for multi-robot obstacle avoidance based on Thiessen polygons, the method comprising the following steps:

[0055] S1. Initialize the finite space environment to obtain the selected bounded region;

[0056] S1.1 Define a rectangular space containing several obstacles and chasing robots to obtain a rectangular bounded region;

[0057] Let the rectangle's length and width be L and W, respectively, and the obstacle's radius be: Number of escape robots and pursuit robots: N E N P .

[0058] S1.2. Obtain the positions of obstacles, chasing robots, and escaping robots within the rectangular bounded area in real time to obtain the layout of the bounded area;

[0059] The location of the obstacle is: The location of the pursuit robot is The location of the escape robot is:

[0060] S1.3. Set the obstacle safety distance and the capture conditions for the escaped robot in the bounded area layout to obtain the selected bounded area.

[0061] Based on the speed of the pursuit robot and its own collision radius, its safe distance is defined as r, and the escape robot E is defined. i The conditions for being captured are:

[0062]

[0063] Where minDist represents a critical distance for determining whether an escaped robot can be captured. Generally, minDist > 2*r.

[0064] S2, such as Figure 3 As shown, the pursuit robot is divided into Thiessen polygons based on a selected bounded region, resulting in Thiessen polygon partitions.

[0065] S2.1 Connect the targets, including the escape-chasing robot, to obtain the Deloitte triangulation diagram;

[0066] S2.2 Calculate the circumcenter of each Delaunay triangle and connect adjacent circumcenters to obtain the Thiessen polygons of the triangulation edge;

[0067] S2.3. Draw perpendicular bisectors of the Thiessen polygons on the edge of the triangulation and intersect them with the selected bounded region to obtain the Thiessen polygon partitions.

[0068] S3. Assign targets to the Thiessen polygon partitions to obtain the pursuit targets and pursuit target points of the pursuit robot.

[0069] The segmentation and boundary properties of the Thiessen polygon are as follows:

[0070] 1. Each Thiessen polygon contains only one data point for a pursuit robot;

[0071] 2. The point within the Tyson polygon is closest to the location of the corresponding pursuit robot.

[0072] 3. The distance from a point on the side of the Tyson polygon to the location of the chasing robot on both sides is equal.

[0073] Assuming that chasing robots i and j are adjacent to each other using the Thiessen polygon, and their adjacent edges are denoted as RidgeEdge(i,j), in the chase-escape game, the physical meaning of the Thiessen polygon is that, for chasing robots at the same speed, the fastest meeting point between a chasing robot and its Thiessen polygon and its adjacent chasing robots is on the intersection of their Thiessen polygons. This is because at any point q on the intersection of their Thiessen polygons, there exists:

[0074]

[0075] For an escape robot, its Tyson polygonal shape represents the approximate time Δt it would take to be successfully captured when it tries to escape in different directions.

[0076] Therefore, the pursuit robot is assigned a pursuit target and a pursuit target point. The pursuit target is the escape robot within the Thiessen polygon partition that has the smallest distance between the escape robot and the pursuit robot. When there are adjacent edges of the Thiessen polygon, the midpoint of the adjacent edge of the Thiessen polygon formed by the pursuit robot and the pursuit target is taken as the pursuit target point; when there are no adjacent edges of the Thiessen polygon, the location of the pursuit target is taken as the pursuit target point. The formula is expressed as follows:

[0077]

[0078] Where RidgeEdge(i,j) are the adjacent edges of the Thiessen polygon. In order to capture the target, N represents the distance between the escaping robot and the pursuing robot. P In order to hunt down the robot swarm, N E A group of escaped robots.

[0079]

[0080] in To track down the target point, MidPoint is the midpoint. The adjacent edges of the Thiessen polygon formed by the pursuing robot and the pursuing target. The location of the target being pursued.

[0081] S4. Divide the selected bounded area, the selected pursuit robot, and obstacles into a safe area;

[0082] S4.1 Connect all obstacles and all pursuing robots to establish the Delone triangle.

[0083] S4.2, such as Figure 4 As shown, based on the Delaunay triangle diagram, each chasing robot is traversed, and its adjacent points are analyzed. When an adjacent point is another chasing robot, it is divided according to the perpendicular bisector of the adjacent edge, similar to a Thiessen polygon. When the adjacent point is an obstacle, it is divided according to the tangent between the adjacent edge and the obstacle circle. The area formed after division is called the quasi-safe area of ​​that robot. Chasing robot P i / E i The quasi-safe region can be represented as a set of points:

[0084]

[0085] Where q is the position of any point within the quasi-safe region, and R k The radius of the obstacle. Location of the obstacle. Let i be the position of robot i. Let N be the position of robot j. O This represents the number of obstacles.

[0086] S4.3, such as Figure 5 As shown, based on the robot's speed and collision radius, we set a safety distance *r*, and perform an erosion operation with a width of *r* around the safety region polygon to obtain its actual safety region. The formula for the safety region is as follows:

[0087]

[0088] Where q is the position of any point within the safe area, and realV i For the safe zone, V i The quasi-safe zone is defined by r, which represents the safe distance.

[0089] S5, RealV for the real-time safe zone of the pursuit robot. i (t) and real-time target tracking points Obtain the directional control pointing point (Front) for the pursuit and obstacle avoidance of the pursuit robot. i (t) represents the point within the safe zone that is closest to the target point, and its formula is as follows:

[0090]

[0091] Front i (t) represents the target point for pursuit direction control, realV i (t) represents the safe zone. To pursue the target location.

[0092] S6, RealV, the real-time safe zone for escaped robots. i (t) and the real-time location of the tracking robot Obtain the directional control point Front for the escape robot's obstacle avoidance. j (t) represents the point within the safe zone that is furthest from the pursuing robot, and its formula is as follows:

[0093]

[0094] Front j (t) represents the escape direction control point, realV i (t) represents the safe zone. To track the robot's location.

[0095] like Figure 2 As shown, a multi-robot obstacle avoidance system based on Thiessen polygons includes:

[0096] The initialization module is used to initialize the finite space environment to obtain the selected bounded region;

[0097] The Thiessen polygon module is used to divide the pursuit robot into Thiessen polygon partitions based on a selected bounded region.

[0098] The task allocation module is used to allocate targets according to the Thiessen polygon partitions, so as to obtain the pursuit targets and pursuit target points of the pursuit robot;

[0099] The safe zone division module is used to divide a selected bounded area, a selected pursuit robot, and obstacles to obtain a safe zone.

[0100] The pursuit direction point module is used to combine the pursuit target point and the safe area to obtain the direction control pointing point for the pursuit robot to pursue and avoid obstacles;

[0101] The escape direction point module is used to combine the position of the pursuit robot and the safe zone to obtain the directional control pointing point for the escape robot to escape and avoid obstacles.

[0102] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0103] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for multi-robot obstacle avoidance based on Thiessen polygons, characterized in that, Includes the following steps: The finite space environment is initialized to obtain the selected bounded region; Based on the selected bounded region, the pursuit robot is divided into Thiessen polygon partitions to obtain Thiessen polygon partitions. Target assignment is performed on the Thiessen polygon partition to obtain the pursuit targets and pursuit target points of the pursuit robot; The selected bounded area, the selected pursuit robot, and the obstacles are divided into regions to obtain a safe area; By combining the target point and the safe area, the directional control pointing point for the pursuit robot to avoid obstacles is obtained; By combining the location of the pursuit robot and the safe zone, the directional control pointing point for the escape robot to avoid obstacles is obtained; The step of dividing the pursuit robot into Thiessen polygon partitions based on a selected bounded region specifically includes: By connecting the target points of the escape-chasing robot, a Deloitte triangulation diagram is obtained; Calculate the circumcenter of each Delaunay triangle and connect adjacent centers to obtain the Thiessen polygons of the triangulation edge; The perpendicular bisectors of the Thiessen polygons at the edge of the triangulation are drawn to intersect the selected bounded region to obtain the Thiessen polygon partitions. The step of dividing the selected bounded area, the selected pursuit robot, and obstacles into a safe area specifically includes: Construct the Deloitte triangle based on obstacles and the chasing robot; Traverse the adjacent points of the pursuit robot, divide the Deloitte triangle, and obtain the quasi-safe area; When the adjacent point is a chasing robot, it is divided according to the perpendicular bisector of the adjacent edge; when the adjacent point is an obstacle, it is divided according to the tangent between its adjacent edge and the obstacle circle. The area formed after the division is the quasi-safe area of ​​the robot. A safe zone is obtained by performing corrosion at a safe distance from the safe zone. The quasi-safe region is expressed by the following formula: in The location of any point within the quasi-safe area, The radius of the obstacle. Location of the obstacle. For robots Location, For robots Location, The number of obstacles; The safe zone is expressed by the following formula: in For any point within the safe zone, For the safe zone, As a quasi-safe zone, For a safe distance.

2. The method for multi-robot pursuit and obstacle avoidance based on Thiessen polygons according to claim 1, characterized in that, The target of the pursuing robot is the escaped robot within the Tyson polygon partition that has the smallest distance between itself and the pursuing robot. Specifically, the target point is: When there are adjacent edges of the Thiessen polygon, the midpoint of the adjacent edges of the Thiessen polygon formed by the pursuing robot and the pursuing target is taken as the pursuing target point. When there are no adjacent edges of the Thiessen polygon, the location of the target to be pursued is taken as the target point.

3. The method for multi-robot pursuit and obstacle avoidance based on Thiessen polygons according to claim 1, characterized in that, The direction control pointing point for the pursuit and obstacle avoidance robot is expressed by the following formula: in To control the direction of pursuit, For the safe zone, To pursue the target location.

4. The method for multi-robot pursuit and obstacle avoidance based on Thiessen polygons according to claim 1, characterized in that, The direction control pointing point for the escape robot to avoid obstacles is expressed by the following formula: in To control the direction of escape, For the safe zone, To track the robot's location.

5. A multi-robot obstacle avoidance system based on Thiessen polygons, characterized in that, A method for performing a multi-robot obstacle avoidance pursuit based on Thiessen polygons as described in claim 1 includes: The initialization module is used to initialize the finite space environment to obtain the selected bounded region; The Thiessen polygon module is used to divide the pursuit robot into Thiessen polygon partitions based on a selected bounded region. The task allocation module is used to allocate targets according to the Thiessen polygon partitions, so as to obtain the pursuit targets and pursuit target points of the pursuit robot; The safe zone division module is used to divide a selected bounded area, a selected pursuit robot, and obstacles into safe zones. The pursuit direction point module is used to combine the pursuit target point and the safe area to obtain the direction control pointing point for the pursuit robot to pursue and avoid obstacles; The escape direction point module is used to combine the position of the pursuit robot and the safe zone to obtain the directional control pointing point for the escape robot to escape and avoid obstacles.

Images