Multi-robot path dynamic margin robust planning method and system based on delay evaluation

Abstract

The invention discloses a multi-robot path dynamic margin robust planning method and system based on delay evaluation. The method comprises the steps that a map and robots are given, and each robot in the map has a starting point and an ending point; setting delay probability and reliability requirements, establishing a global timetable for each node, and initializing the global timetable to be blank; allocating a priority of path planning for each robot in the map; and sequentially selecting one robot as the current to-be-planned robot according to the priority order until all the robots are planned. According to the method, the delay conflict constraint interval is dynamically adjusted in the planning process, the conflict risk possibly generated in path execution is effectively reduced, a dynamic adjustment strategy is adopted for the robust coefficient to ensure that the robot path meets the expected reliability requirement, and the problem of insufficient reliability caused by high path coupling in the prior art is solved. It is further shown through simulation that the reliability requirement can be improved on the premise of the same path time cost, the solving speed is remarkably improved under the high reliability requirement, and therefore the calculation cost is reduced.

Description

Technical Field

[0001] The invention relates to a multi-robot path dynamic margin robust planning method and system based on delay evaluation, belonging to the field of multi-robot path planning. Background Art

[0002] Multi-robot path planning has a large number of engineering applications in warehousing logistics, unmanned driving systems, airport transportation and other fields. The problem is to find conflict-free paths for multiple robots with given starting and ending points.

[0003] In the actual execution process of a multi-robot system, various uncertain factors may affect the actual execution path and the planned path, which may cause conflicts between robots and cause algorithm execution failure. When various unexpected events occur, it may be necessary to adjust the plans of one or more robots to avoid collisions. However, re-planning may require a lot of computing and communication capabilities, and it is difficult to ensure that the planning is completed before the robots collide.

[0004] At present, the k-robust planning algorithm is mainly used to ensure the execution reliability of multi-robot systems. However, the delay in the system is often difficult to determine, which makes it difficult to determine the appropriate k to deal with different scenarios. If k is set too large, efficiency and energy saving will be sacrificed in exchange for reliability. If k is set too small, it is difficult to ensure stable and reliable execution of the system. Therefore, it is necessary to consider a reliable and efficient robust path planning method. Summary of the invention

[0005] The present invention provides a multi-robot path dynamic margin robust planning method based on delay evaluation. Based on the delay probability and combined with the single robot execution reliability requirement, a safety interval planning method is adopted. Under the safety constraints and delay conflict constraints, paths that meet the reliability requirements are planned for all robots in order of priority.

[0006] The technical solution of the present invention is:

[0007] According to a first aspect of the present invention, a multi-robot path dynamic margin robust planning method based on delay evaluation is provided, comprising the following steps:

[0008] Step 1: Given a map and a robot, each robot in the map has a starting point and an end point;

[0009] Step 2: Set the delay probability and reliability requirements, create a global schedule for each node and initialize it to blank;

[0010] Step 3: Assign path planning priorities to each robot in the map; select one robot in order of priority as the current robot to be planned until all robots have completed planning:

[0011] If the current robot to be planned is the highest priority, the path planning algorithm is directly used to search and expand the nodes to find a path from the starting point to the end point to achieve planning; and the safety constraint interval and delay conflict constraint interval generated by each node on the path of the highest priority robot are inserted into the global schedule. The intervals other than the safety constraint interval and the delay conflict constraint interval are safety intervals; among them, the delay conflict constraint interval is inserted sequentially after the safety constraint interval of the global schedule;

[0012] Otherwise, if the current robot to be planned is not the highest priority, then according to the global schedule obtained after the previous planning of the current robot to be planned, the path planning algorithm is used to search and expand the nodes according to the safety interval, and a path from the starting point to the end point is found to realize the planning; and the safety constraint interval and delay conflict constraint interval generated by each node on the path of the currently planned robot are inserted into the global schedule, and the intervals other than the safety constraint interval and the delay conflict constraint interval are safety intervals; among which, the delay conflict constraint intervals are inserted sequentially after the safety constraint intervals of the global schedule.

[0013] Furthermore, in step 3, the safety constraint interval generated by each node is the time step of the robot passing through the node, and the delay conflict constraint interval generated by each node is specifically: the robust coefficient k is incremented from the starting robust coefficient by a preset step length and P(p i (t),k)) until P(p i (t),k)) is greater than the reliability requirement R, which will satisfy “P(p i The minimum robustness coefficient k when (t),k)) is greater than the reliability requirement R” is taken as the robot R i The delay conflict constraint interval at the current node.

[0014] Furthermore, the specific discriminant expression of the delay conflict constraint interval generated by each node is as follows:

[0015]

[0016] Among them, P(p i (t),k)) indicates that the robot starts from the starting point p i (0) Execute to the current node p i (t), the probability of a delay within k time steps; n represents the robot R i From the starting point p i (0) to the current node p i (t) is the delay time step, m represents the strict execution time step; T is the maximum time step; p d is the delay probability.

[0017] According to a second aspect of the present invention, there is provided a multi-robot path dynamic margin robust planning system based on delay evaluation, comprising a module of any one of the methods described above.

[0018] According to a third aspect of the present invention, a processor is provided, wherein the processor is used to run a program, wherein when the program is run, any one of the above-mentioned multi-robot path dynamic margin robust planning methods based on delay evaluation is executed.

[0019] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, which includes a stored program, which, when executed by a processor, enables the processor to implement any of the above-mentioned multi-robot path dynamic margin robust planning methods based on delay evaluation.

[0020] The beneficial effects of the present invention are as follows: the present invention dynamically adjusts the delay conflict constraint interval during the planning process, effectively reducing the conflict risk that may occur during path execution, and adopts a dynamic adjustment strategy for the robustness coefficient k to ensure that the robot path meets the expected reliability requirements, overcoming the problem of insufficient reliability caused by high path coupling in the prior art. Further simulations show that the present invention can improve the reliability requirements under the premise of the same path time cost, significantly improve the solution speed under high reliability requirements, thereby reducing the computing cost, and providing a more robust path planning solution for the application of multi-robot collaborative systems in complex scenarios such as logistics and autonomous driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a flow chart of the present invention;

[0022] Figure 2 are three robot path planning maps provided by embodiments of the present invention;

[0023] Figure 3 are three robot path planning time diagrams provided by embodiments of the present invention;

[0024] Figure 4 is an example of a global timetable after inserting a barrier interval provided by an embodiment of the present invention;

[0025] Figure 5 This is a graph showing the cost-reliability comparison experiment results of the present invention and other algorithms on a 32×32-20 random map;

[0026] Figure 6 This is a graph showing the cost-reliability comparison experimental results between the present invention and other algorithms on a 48×48 empty map. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. It should be noted that the embodiments in this application and the features in the embodiments can be combined with each other arbitrarily without conflict.

[0028] Example 1: Figure 1-Figure 6 As shown, according to a first aspect of an embodiment of the present invention, a multi-robot path dynamic margin robust planning method based on delay evaluation is provided, comprising the following steps:

[0029] Step 1: Given a map (G = V, E), V represents the set of nodes in the map, E represents the set of edges in the map, and a multi-robot set U = {R1, R2, ... R n}, each robot has a starting point and an end point.

[0030] For example: Figure 2 As shown, in the 6×4 grid map, there are 24 nodes in total, and the three robots are numbered R1, R2, and R3 respectively; Figure 2 S1, S2, and S3 are the grids where the three robots’ respective starting points are located, and G1, G2, and G3 are the grids where the three robots’ respective end points are located. In the figure, the black grids are inaccessible obstacle areas, and the white grids are passable areas. It is known that at the initial time t=0, each robot is at the starting point.

[0031] Step 2: Set the delay probability and reliability requirements, create a global schedule for each node and initialize it to blank;

[0032] For example, before planning a path, a group of robots sets a delay probability p d =0.1, single robot reliability requirement R = 98%; establish a global schedule for each node, the time of the schedule is from t = 0, 1, ... T; in the embodiment of the present invention, T is 9, and can also be set to other values ​​as needed, each time step is 1, and the time represented by the time step is defined as needed.

[0033] Step 3: Assign path planning priorities to each robot in the map; select one robot in order of priority as the current robot to be planned until all robots have completed planning:

[0034] If the current robot to be planned is the highest priority, the SIPP algorithm is directly used to search and expand nodes to find a path from the starting point to the end point to achieve planning; and the safety constraint interval and delay conflict constraint interval generated by each node on the path of the highest priority robot are inserted into the global schedule. The intervals other than the safety constraint interval and the delay conflict constraint interval are safety intervals; among them, the delay conflict constraint interval is inserted sequentially after the safety constraint interval of the global schedule;

[0035] Otherwise, if the current robot to be planned is not the highest priority, then according to the global schedule obtained after the previous planning of the current robot to be planned, the SIPP algorithm is used to search and expand the nodes according to the safety interval, and a path from the starting point to the end point is found to realize the planning; and the safety constraint interval and delay conflict constraint interval generated by each node on the path of the currently planned robot are inserted into the global schedule, and the interval except the safety constraint interval and the delay conflict constraint interval is the safety interval; wherein, the delay conflict constraint interval is sequentially inserted after the safety constraint interval of the global schedule; the SIPP algorithm is an existing well-known safety interval planning algorithm, and the SIPP algorithm in the present invention can also be replaced by other similar planning algorithms.

[0036] Furthermore, in step 3, the safety constraint interval generated by each node is the time step of the robot passing through the node, and the delay conflict constraint interval generated by each node is specifically: the robust coefficient k is incremented from the starting robust coefficient by a preset step length and P(p i (t),k)) until P(p i (t),k)) is greater than the reliability requirement R, which means that the current robustness coefficient k can plan a path with a reliability of at least R for a single robot, that is, the robot R i The probability of conflicts with other robots due to various types of non-deterministic delay disturbances will not be greater than (1-R), which will satisfy "P(p i The minimum robustness coefficient k when (t),k)) is greater than the reliability requirement R” is taken as the robot R i In the delay conflict constraint interval of the current node. In the embodiment of the present invention, the preset step size is set to 1.

[0037] Furthermore, the specific discriminant expression of the delay conflict constraint interval generated by each node is as follows:

[0038]

[0039] Among them, P(p i (t),k)) indicates that the robot executes to the current node p i (t), the probability of a delay within k time steps; n represents the robot Ri From the starting point p i (0) to the current node p i (t) is the delay time step, m represents the strict execution time step; T is the maximum time step; p d is the delay probability; k is the robot R i The robustness coefficient at the current node; R is the reliability requirement.

[0040] This example aims to find a path from the starting point to the end point for each robot, and any two paths satisfy the safety constraints and delay conflict constraints, that is, a conflict-free path can be obtained when a delay disturbance within the robustness coefficient k occurs between any two robots. The overall path plan obtained by repeating the above-mentioned single-machine planning process until all robots have completed the planning is the conflict-free path planning plan of the multi-robot path dynamic margin robust planning method based on delay evaluation. The global path plan finally generated by the present invention optimizes the overall path time cost and execution efficiency while meeting the reliability requirements.

[0041] According to the second aspect of an embodiment of the present invention, there is provided a multi-robot path dynamic margin robust planning system based on delay evaluation, including a module of any of the methods described above. Specifically comprising: a first module, for a given map, in which each robot has a starting point and an end point; a second module, for setting delay probability and reliability requirements, establishing a global schedule for each node and initializing it to blank; a third module, for assigning path planning priorities to each robot in the map; selecting one robot in turn in order of priority as the current robot to be planned, until all robots have completed planning. For the parts not described in detail in the above for each module, please refer to the relevant description of the embodiment.

[0042] According to a third aspect of an embodiment of the present invention, a processor is provided, wherein the processor is used to run a program, wherein when the program is running, any one of the above-mentioned multi-robot path dynamic margin robust planning methods based on delay evaluation is executed.

[0043] According to a fourth aspect of an embodiment of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, which, when executed by a processor, enables the processor to implement any of the above-mentioned multi-robot path dynamic margin robust planning methods based on delay evaluation.

[0044] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0045] The following is an example of three machines:

[0046] 1. Assign path planning priorities to each robot in the map. The priority numbers of R1, R2, and R3 are 1, 2, and 3. The smaller the number, the higher the priority. The robot priorities are R1>R2>R3. Therefore, the path planning of the highest priority robot R1 is performed first, and the path of robot R1 is obtained as shown in Table 1:

[0047] Table 1

[0048] time t＝0 t＝1 t＝2 t＝3 t＝4 t＝5 t＝6 node (A,3) (A,2) (B,2) (C,2) (D,2) (E,2) (F,2)

[0049] 2. Path Planning

[0050] In this embodiment, the time step required for the robot to enter a node and exit the node is set to 1. The initial robustness coefficient is 1, and the preset step length is 1.

[0051] Table 2 shows the obstacle interval of the robot R1 path (the obstacle interval is the safety constraint interval + the delay conflict constraint interval). At t = 0, the probability that the robot R1 executes to the current node (A, 3) and there is a delay of less than 1 time step exceeds 98%. In this time interval, considering the robustness coefficient k = 1, the reliability requirement R = 98% can be met. Then, t = 0 is the safety constraint interval, and t = 1 is the delay conflict constraint interval (that is, the delay conflict constraint interval is inserted sequentially after the safety constraint interval, t = 0 is the safety constraint interval, k = 1, that is, the delay conflict constraint interval is 1 time step after the safety constraint interval); t = [0 ,1] constitutes the obstacle interval of the position point (A,3), that is, when the subsequent robot expands to the (A,3) node, it needs to search for a path in the time interval outside the obstacle interval constraint of t=[0,1]; similarly, at t=5, considering the robustness coefficient k=1, it can no longer meet the reliability requirement R=98%, so considering the robustness coefficient k=2, t=5 is the safety constraint interval, t=[6,7] is the delayed conflict constraint interval, and t=[5,7] constitutes the obstacle interval of the position point (E,2); the other nodes are understood in the same way; the obstacle interval of the robot R1 is finally obtained as shown in the third column of Table 2.

[0052] Table 2

[0053] time node Obstacle area t＝0 (A,3) t＝[0,1] t＝1 (A,2) t＝[1,2] t＝2 (B,2) t＝[2,3] t＝3 (C,2) t＝[3,5] t＝4 (D,2) t＝[4,6] t＝5 (E,2) t＝[5,7] t＝6 (F,2) t＝[6,8]

[0054] After the planning of robot R1 is completed, the next priority robot R2 is used as the current robot to be planned. The path of R2 is found within the time interval outside the obstacle interval constraint. The SIPP algorithm is used to search and expand the nodes to find a path from the starting point to the end point to realize the planning; then the planning of R3 is realized.

[0055] According to Figure 2 The map shown in the figure, applying the above technical solution, obtains the path planning result as follows Figure 3 shown.

[0056] In the above path planning process, there is a conflict at the two nodes (B, 2) and (E, 2). For a better understanding, the division method of the safety constraint interval and the delay conflict constraint interval proposed by the present invention is described in conjunction with the accompanying drawings:

[0057] like Figure 4 As shown, robot R1 occupies position point (B, 2) at t = 2, at which time the robustness coefficient k of R1 is 1, and t = [2, 3] becomes the obstacle interval of position point (B, 2); robot R2 does not occupy this position point, and robot R3, which has a priority after robot R2, can only enter (B, 2) in the safety interval t = [4, ∞] of position point (B, 2) to reserve a time margin of 1 time step between robots to cope with unexpected delays; similarly, robot R1 occupies position point (E, 2) at t = 5, and the robustness coefficient k at this time is 2. At the same time, robot R2 occupies this position point on the path assigned to it. To ensure reliability, robot R2 can only enter (E, 2) in the safety interval t = [8, ∞]. Figure 4 From top to bottom, they are: the global timetable (B, 2) obtained by robot R1 after the planned path is completed, the global timetable (B, 2) obtained by robot R3 after the planned path is completed, the global timetable (E, 2) obtained by robot R1 after the planned path is completed, and the global timetable (E, 2) obtained by robot R2 after the planned path is completed.

[0058] 3. Verification Instructions

[0059] 1) Select a typical random map 32×32 map, use the execution time as the termination condition, set the maximum time limit for a single solution to 120 seconds, and if the time limit is exceeded, it is considered a solution failure. In the test, the AkR-SIPP algorithm proposed in the present invention sets the reliability requirements of a single robot to 85%, 90%, and 95%, respectively. The corresponding starting robustness coefficient k={1,2,3} under each reliability requirement, and the existing kR-SIPP algorithm uses a robustness coefficient k={1,2,3,4,5,6,7,8,9}. The robustness coefficients of the two algorithms are incremented with a step size of 1. Randomly select 60 and 75 robots to conduct 10,000 experiments based on the above parameters, respectively, to solve the cost-reliability performance of robots of different sizes. Figure 5 As shown in (cost value is path length), under the condition of the same number of robots, the average running time and average maximum completion time of each algorithm in different reliability intervals are compared as shown in Table 3 (running time is the time it takes for multiple robots to complete path planning, and the average running time is the average of multiple experiments; maximum completion time is the maximum completion time for multiple robots to run from the starting point to the end point according to the planned path, and the average maximum completion time is the average of multiple experiments; M is the overall execution reliability index of the system obtained after the path plan is verified by the Monte Carlo method).

[0060] Table 3

[0061]

[0062] 2) Select a typical empty map 48*48 map, use the execution time as the termination condition, set the maximum time limit for a single solution to 120 seconds, and if the time limit is exceeded, it is considered a solution failure. In the test, the reliability requirements of the AkR-SIPP algorithm proposed in the present invention are set to 85%, 90%, and 95%, respectively, and the initial robustness coefficient k = {1, 2, 3}. The existing kR-SIPP algorithm uses a robustness coefficient k = {1, 2, 3, 4, 5, 6, 7, 8, 9}. The robustness coefficients of the two algorithms are increased in steps of 1. 90 and 130 robots are randomly selected and 10,000 experiments are performed based on the above parameters to solve the cost-reliability performance of robots of different sizes. Figure 6As shown, under the condition of the same number of robots, the average running time and average maximum completion time of each algorithm in different reliability intervals are compared as shown in Table 4 (the running time is the time for multiple robots to complete path planning, and the average running time is the average value of multiple experiments; the maximum completion time is the maximum completion time for multiple robots to run from the starting point to the end point according to the planned path, and the average maximum completion time is the average value of multiple experiments). The results show that the cost of the method of the present invention can be reduced by 25% compared with kR-SIPP in the same high reliability interval. At the same time, the solution speed can be improved by 45% in the same high reliability interval (that is, the overall reliability M of the system is required to be above 33%). Especially in large-scale robot planning, the average running time of the present invention in the high reliability interval is more significantly improved. Since the maximum completion time may be affected by many factors under real conditions, and the present invention mainly studies the running time, the main indicator is still the average running time of path planning. As Figure 6 As shown in Table 4, it is shown that the present invention provides a more efficient and reliable robust solution.

[0063] Table 4

[0064]

[0065] The specific implementation modes of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above implementation modes, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A multi-robot path dynamic margin robust planning method based on delayed evaluation, characterized in that: The following steps are involved: Step 1: Given a map and a robot, each robot in the map has a starting point and an end point; Step 2: Set the delay probability and reliability requirements, create a global schedule for each node and initialize it to blank; Step 3: Assign path planning priorities to each robot in the map; select one robot in order of priority as the current robot to be planned until all robots have completed planning: If the current robot to be planned is the highest priority, the path planning algorithm is directly used to search and expand the nodes to find a path from the starting point to the end point to achieve planning; and the safety constraint interval and delay conflict constraint interval generated by each node on the path of the highest priority robot are inserted into the global schedule. The intervals other than the safety constraint interval and the delay conflict constraint interval are safety intervals; among them, the delay conflict constraint interval is inserted sequentially after the safety constraint interval of the global schedule; Otherwise, if the current robot to be planned is not the highest priority, then according to the global schedule obtained after the previous planning of the current robot to be planned, the path planning algorithm is used to search and expand the nodes according to the safety interval, and a path from the starting point to the end point is found to realize the planning; and the safety constraint interval and delay conflict constraint interval generated by each node on the path of the currently planned robot are inserted into the global schedule, and the intervals other than the safety constraint interval and the delay conflict constraint interval are safety intervals; among which, the delay conflict constraint intervals are inserted sequentially after the safety constraint intervals of the global schedule.

2. The multi-robot path dynamic margin robust planning method based on delay evaluation according to claim 1 is characterized in that: In step 3, the safety constraint interval generated by each node is the time step of the robot passing through the node. The delay conflict constraint interval generated by each node is specifically: the robust coefficient k is incremented from the starting robust coefficient by a preset step length and P(p i (t),k)) until P(p i (t),k)) is greater than the reliability requirement R, which will satisfy "P(p i The minimum robustness coefficient k when (t),k)) is greater than the reliability requirement R” is taken as the robot R i The delay conflict constraint interval at the current node.

3. The multi-robot path dynamic margin robust planning method based on delay evaluation according to claim 2 is characterized in that: The specific discriminant expression of the delay conflict constraint interval generated by each node is as follows: Among them, P(p i (t),k)) indicates that the robot starts from the starting point p i (0) Execute to the current node p i (t), the probability of a delay within k time steps; n represents the robot R i From the starting point p i (0) to the current node p i (t) is the delay time step, m represents the strict execution time step; T is the maximum time step; p d is the delay probability.

4. A multi-robot path dynamic margin robust planning system based on delayed evaluation, characterized in that: A module comprising the method according to any one of claims 1 to 3.

5. A processor, characterized in that: The processor is used to run a program, wherein the program, when running, executes the multi-robot path dynamic margin robust planning method based on delay evaluation as described in any one of claims 1 to 3.

6. A computer-readable storage medium, the computer-readable storage medium comprising a stored program, characterized in that: When the program is executed by a processor, the processor implements the multi-robot path dynamic margin robust planning method based on delay evaluation as described in any one of claims 1 to 3.

Images