Robot global path planning method and system based on probability interval division

Abstract

The present invention proposes a robot global path planning method and system based on probability interval division, and relates to the field of robot path planning technology. It includes calculating the sampling probability of the obstacle-based interval based on the obstacle area and the total area of ​​obstacles in each equally divided interval; determining the positions of the starting point and the end point, and the equally divided interval containing the starting point and the end point, and setting the sampling probability of the position-based interval between the equally divided intervals containing the starting point and the end point; combining the sampling probability of the obstacle-based interval with the sampling probability of the position-based interval to obtain the sampling probability in the region; based on the RRT* algorithm, applying the sampling probability in the region to select nodes, sampling the equally divided intervals in a variable step size manner, and iteratively generating the initial path; optimizing the generated initial path to obtain the global path of the robot. The present invention can accelerate the speed of generating preliminary paths by growing random trees, and enhance the purposefulness and directionality of the algorithm.

Description

Technical Field

[0001] The present invention belongs to the technical field of robot global path planning, and in particular relates to a robot global path planning method and system based on probability interval division. Background Art

[0002] In recent years, with the continuous advancement of hardware, robots have been widely used in many fields such as medicine, military, industry, agriculture and service industry. Robot path planning refers to finding an optimal or suboptimal collision-free path from the starting point to the end point under certain evaluation criteria. Many mature algorithms have been proposed to solve this problem, among which the RRT (rapid exploration tree based on sampling) algorithm performs well in high-dimensional and multi-constrained path planning, especially suitable for single query. Commonly used algorithms for local path planning include DWA (dynamic window method) and APF (artificial potential field method).

[0003] RRT and its variants have advantages such as simple principle, probability completeness, strong search capability, and wide application, but they also have disadvantages such as long planning time, high path cost, and rough curves. In addition, RRT algorithms are usually not purposeful, and they do not consider the impact of obstacle size on the path.

[0004] Since the traditional Informed-RRT* (Information-enhanced Rapid Random Trees) algorithm is based on random sampling, the position and direction of its new growth points are random, so the path it finally generates deviates from the optimal path, and a large number of samples are required when planning paths in narrow areas, which seriously affects the planning efficiency. Summary of the invention

[0005] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides a robot global path planning method and system based on probability interval division, introduces random sampling within the probability interval, dynamically constrains the probability of sampling points, divides the map into equal areas, and calculates the sampling probabilities of different areas based on the plane area occupied by obstacles and the locations of the starting point and the end point, thereby accelerating the speed of growing random trees to generate preliminary paths and enhancing the purposefulness and directionality of the Informed-RRT* algorithm.

[0006] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0007] A first aspect of the present invention provides a robot global path planning method based on probability interval division.

[0008] The robot global path planning method based on probability interval division includes the following steps:

[0009] Step 1: Obtain the total map area of ​​the robot's global path planning, divide the total map area into N equal parts, and count the obstacle area in each divided interval and the total obstacle area;

[0010] Step 2: Based on the obstacle area in each equally divided interval and the total obstacle area, calculate the sampling probability in the interval based on the obstacle;

[0011] Step 3: Determine the positions of the starting point and the end point, and the equally divided intervals including the starting point and the end point, and set the sampling probability within the interval based on the position of the equally divided intervals including the starting point and the end point;

[0012] Step 4: Combine the sampling probability within the interval based on obstacles and the sampling probability within the interval based on positions to obtain the sampling probability within the area;

[0013] Step 5: Based on the RRT* algorithm, apply the sampling probability in the area to select nodes, and find the node closest to the selected node. Use the variable step size method to sample the equally divided intervals to obtain new nodes extending from the node closest to the selected node to the selected node, and connect the new node and the node closest to the selected node to obtain the current path tree;

[0014] Step 6: Iterate the above step 5 until both the starting point and the target point are included in the path tree, stop the iteration, and obtain the initial path;

[0015] Step 7: Optimize the generated initial path to obtain the global path of the robot.

[0016] A second aspect of the present invention provides a robot global path planning system based on probability interval division.

[0017] The robot global path planning system based on probability interval division includes:

[0018] The area statistics module is configured to: obtain the total map area of ​​the robot's global path planning, divide the total map area into N equal parts, and count the obstacle area in each equal division and the total obstacle area;

[0019] The first sampling probability calculation module is configured to: calculate the sampling probability within the obstacle-based interval based on the obstacle area and the total obstacle area within each equally divided interval;

[0020] The second sampling probability calculation module is configured to: determine the positions of the starting point and the end point, and the equally divided interval including the starting point and the end point, and set the sampling probability within the interval based on the position of the equally divided interval including the starting point and the end point;

[0021] The total sampling probability calculation module is configured to: combine the sampling probability within the interval based on obstacles and the sampling probability within the interval based on positions to obtain the sampling probability within the area;

[0022] The initial path generation module is configured as follows: based on the RRT* algorithm, the sampling probability in the region is applied to select nodes, and the node closest to the selected node is found, and the equally divided interval is sampled in a variable step size manner to obtain a new node extending from the node closest to the selected node to the selected node, and the new node is connected to the node closest to the selected node to obtain the current path tree;

[0023] The iteration module is configured to: iterate the initial path generation module until the starting point and the target point are both included in the path tree, stop the iteration, and obtain the initial path;

[0024] The path optimization module is configured to optimize the generated initial path to obtain the global path of the robot.

[0025] One or more of the above technical solutions have the following beneficial effects:

[0026] The present invention discloses a robot global path planning method and system based on probability interval division, which introduces random sampling in probability intervals, dynamically constrains the probability of sampling points, divides the map into equal areas, and gives different sampling probabilities to different areas according to the inverse ratio of the area of ​​each obstacle to the total obstacles and the positions of the starting point and the end point. The sampling probability in the area is obtained by considering the sampling probability in the interval based on obstacles and the sampling probability in the interval based on positions, so as to accelerate the speed of generating preliminary paths by growing random trees and enhance the purposefulness and directionality of the Informed-RRT* algorithm.

[0027] When calculating the sampling probability within an obstacle-based interval, the present invention performs calculations in inverse proportion to the proportion of each obstacle to the total obstacle area, and increases the sampling rate rather than reduces it in areas with fewer obstacles. From an exploratory perspective, areas with fewer obstacles can be explored more comprehensively, which helps to discover potential paths; from a safety perspective, the risk of collision is reduced; from an efficiency perspective, feasible paths can be found more quickly, improving path planning efficiency; from a target proximity perspective, increasing the sampling rate in areas with fewer obstacles can find the path to the target more quickly.

[0028] When calculating the final sampling probability within the area, the two probability weights are dynamically adjusted accordingly, and the sampling probability within the interval based on obstacles is used as the reference value of the weight to avoid too many obstacles around the starting point and the end point, while there are fewer obstacles at other locations.

[0029] The present invention adopts a variable step length scheme to generate an initial path, and optimizes the generated initial path to finally obtain a global path of the robot.

[0030] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0032] Figure 1 This is a flow chart of the method of the first embodiment.

[0033] Figure 2 The global path map of the robot generated using the traditional Informed-RRT* method.

[0034] Figure 3 This is a global path map of the robot generated using the method of the first embodiment. DETAILED DESCRIPTION

[0035] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. 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 invention belongs.

[0036] It should be noted that the terms used herein are for describing specific embodiments only and are not intended to be limiting of exemplary embodiments according to the present invention.

[0037] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0038] Embodiment 1

[0039] This embodiment discloses a robot global path planning method based on probability interval division.

[0040] like Figure 1 As shown, the robot global path planning method based on probability interval division includes the following steps:

[0041] Step 1: Obtain the total map area of ​​the robot's global path planning, divide the total map area into N equal parts, and count the obstacle area in each divided interval and the total obstacle area;

[0042] Step 2: Based on the obstacle area in each equally divided interval and the total obstacle area, calculate the sampling probability in the interval based on the obstacle;

[0043] Step 3: Determine the positions of the starting point and the end point, and the equally divided intervals including the starting point and the end point, and set the sampling probability within the interval based on the position of the equally divided intervals including the starting point and the end point;

[0044] Step 4: Combine the sampling probability within the interval based on obstacles and the sampling probability within the interval based on positions to obtain the sampling probability within the area;

[0045] Step 5: Based on the RRT* algorithm, apply the sampling probability in the area to select nodes, and find the node closest to the selected node. Use the variable step size method to sample the equally divided intervals to obtain new nodes extending from the node closest to the selected node to the selected node, and connect the new node and the node closest to the selected node to obtain the current path tree;

[0046] Step 6: Iterate the above step 5 until both the starting point and the target point are included in the path tree, stop the iteration, and obtain the initial path;

[0047] Step 7: Optimize the generated initial path to obtain the global path of the robot.

[0048] The embodiment of the present invention introduces random sampling within a probability interval, dynamically constrains the probability of sampling points, divides the map into equal areas, and gives different sampling probabilities to different areas according to the plane area occupied by obstacles and the locations of the starting point and the end point, thereby accelerating the speed of generating preliminary paths by growing random trees and enhancing the purposefulness and directionality of the Informed-RRT* algorithm.

[0049] In some embodiments:

[0050] The step five specifically includes:

[0051] Set multiple parameters of the RRT* algorithm and randomly select a node in the map using the obtained sampling probability within the region ;

[0052] Find the distance in the initial tree The nearest node ;

[0053] calculate and Distance , using a greedy step size adjustment strategy based on distance Adjust the step length ;

[0054] from Orientation Expand and generate new nodes ;

[0055] judge Is it a valid node without collision? If so, Add to the tree and connect and ;

[0056] Cycle the above process to find the first initial path from the starting point to the target point .

[0057] Next, we will combine Figure 1 , elaborate on the technical solution of this embodiment. The detailed steps of the algorithm implementation of this embodiment may include:

[0058] 1. Calculating the sampling probability in the area

[0059] 1. Get the total map area of ​​the robot's global path planning. The total map area is recorded as ;

[0060] The total area of ​​the map is divided into N equal parts, and the divided intervals are marked as , , , … ; is the total number of equally divided intervals.

[0061] 2. Count the obstacle area in each divided area and the total obstacle area:

[0062] Above The areas of obstacles in the equally divided intervals are recorded as , Indicates The total area of ​​obstacles is .

[0063] 3. Based on the obstacle area in each equally divided interval and the total obstacle area, calculate the sampling probability within the interval based on the obstacle.

[0064] Among them, the specific formula for the sampling probability in the area based on obstacles is:

[0065] .

[0066] in, is the sampling probability within the interval based on obstacles, Indicates Equally divided intervals; For the The obstacle area between the equal partitions; is the total number of equally divided intervals; The cumulative sum is Obstacle area.

[0067] The sampling probability in the obstacle-based area is inversely proportional to the proportion of each obstacle to the total obstacles, and is normalized.

[0068] In this embodiment, the sampling rate is increased instead of reduced in areas with fewer obstacles, based on the following considerations:

[0069] Exploratory:

[0070] Increasing the sampling rate in areas with fewer obstacles allows those areas to be explored more thoroughly, which helps discover potential paths and optimize routes. If you reduce the sampling rate, some feasible paths may be missed.

[0071] Security:

[0072] Areas with fewer obstacles generally mean more free space, which reduces the risk of collisions. Therefore, increasing the sampling rate allows these areas to be explored more safely.

[0073] efficiency:

[0074] In areas with fewer obstacles, the path planning algorithm can find a feasible path more quickly. Increasing the sampling rate can further improve this efficiency because it can find a path to the goal more quickly.

[0075] Target proximity:

[0076] If the target is located in an area with fewer obstacles, increasing the sampling rate in these areas can help find a path to the target faster.

[0077] 4. Determine the positions of the start and end points, as well as the equally divided intervals containing the start and end points, and set the position-based sampling probability within the equally divided intervals containing the start and end points.

[0078] The rule for determining the sampling probability within the interval based on the position is:

[0079] First, determine the area where the starting point start is located, and set the sampling probability based on the position in the interval between the equal partitions where the starting point start is located, denoted as η1=0.5;

[0080] Determine the area where the goal is located, and set the sampling probability based on the position in the interval between the equal partitions where the goal is located, denoted as η2=0.5;

[0081] The rest =0.

[0082] In this embodiment, the sampling probability within the position-based interval between the equal partitions including the starting point and the end point is set mainly based on optimization goal orientation.

[0083] Optimized goal-oriented: By increasing the sampling rate of the start and end points, the algorithm can find the path to the goal faster. This helps ensure the success rate of the path planning task, especially when there is a time limit or the destination point is important.

[0084] 5. Combine the sampling probability within the interval based on obstacles and the sampling probability within the interval based on location to obtain the sampling probability within the area.

[0085] Determine the sampling probability within the area. The specific formula is:

[0086] .

[0087] in, For the The sampling probability within the region between the equal partitions; For the The probability of sampling within the obstacle-based interval between equal partitions; For the The sampling probability of the interval based on the position between the equal partitions; and is the probability weight.

[0088] In order to improve the robustness of the algorithm and avoid the situation where there are too many obstacles around the starting point and the end point, but few obstacles at other locations, it is necessary to weight the two probabilities. and Make corresponding dynamic adjustments.

[0089] Therefore, the sampling probability within the former obstacle interval is used as the reference value of the weight:

[0090] ;

[0091] .

[0092] but:

[0093] .

[0094] 6. Coordinate generation:

[0095] (1) Generate random numbers :

[0096] .

[0097] is a random number between 0 and 1.

[0098] (2) Calculate the cumulative sampling probability :

[0099] .

[0100] in, Indicates to The cumulative sampling probability until the interval is divided into equal parts.

[0101] (3) Select interval:

[0102] According to random number Select the interval :if ,choose ;if ,choose ;if ,choose ;if , choose .

[0103] (4) In the selected equal division interval Generate a random coordinates and a random This can be achieved by generating two random numbers, one at of Range and Assume of The range is from arrive , The range is from arrive , then you can generate random points as follows:

[0104] .

[0105] in, Indicates that in the interval Generate a random number within.

[0106] Therefore, this embodiment uses obstacles and target point positions as judgment nodes and as consideration conditions for generating probability. When generating random points, random points are not generated blindly, but are obtained based on the calculated sampling probability within the area + the sampling probability within the position-based interval set between equal partitions including the starting point and the end point.

[0107] (II) Generate initial path

[0108] When generating the initial path, this embodiment:

[0109] Based on the RRT* algorithm, the sampling probability in the area is applied to select nodes, and the node closest to the selected node is found. The equally divided interval is sampled using a variable step size method to obtain a new node extending from the node closest to the selected node to the selected node, and the new node is connected to the node closest to the selected node to obtain the current path tree;

[0110] The above process is iterated until both the starting point and the target point are included in the path tree, then the iteration is stopped to obtain the initial path.

[0111] This embodiment adopts the above variable step length scheme to generate the initial path, the purpose is:

[0112] Environmental characteristics:

[0113] The initial step size is determined based on the size and complexity of the search space.

[0114] For example, in larger or more complex environments, the initial step size can be set larger to cover the search space faster.

[0115] Target area distance:

[0116] Consider the distance between the starting point and the target area. The initial step size should be large enough to quickly approach the target area in the early stages.

[0117] Obstacle density:

[0118] In environments with more obstacles, smaller step sizes can help avoid collisions, but may require more iterations.

[0119] The variable step length scheme is used to generate the initial path. The specific process is as follows:

[0120] 1. Define the initial tree Contains the start node ;

[0121] Define the target area ;

[0122] Set the initial step size ;

[0123] and yes The minimum and maximum values ​​of

[0124] and yes The minimum and maximum values ​​of

[0125] is the current iteration number;

[0126] is the maximum number of iterations;

[0127] is the average length of the paths in the current tree;

[0128] is the initial average path length (usually it can be set to the initial step length, here it is taken as the benchmark);

[0129] is the expected path length to the target area (of course, the line segment between two points is expected to be the shortest).

[0130] 2. From free space In the above example, a node is randomly selected using the sampling probability method within the region. .

[0131] 3. Select the nearest node:

[0132] In the tree Find the The nearest node :

[0133] .

[0134] 4. Greedy step size adjustment:

[0135] According to distance Adjust step size ,in, and is the parameter that controls the step size adjustment, specifically:

[0136] ;

[0137] ;

[0138] ;

[0139] .

[0140] Adjustments:

[0141] As the number of iterations increases, The value of gradually decreases. At the same time, as the average path length approaches the target path length, This ensures that the step size is large in the initial stages of the algorithm to quickly cover the space, and decreases as the goal is approached to improve accuracy.

[0142] Adjustments:

[0143] As the number of iterations increases, At the same time, as the average path length moves away from the initial path length, This ensures that the step size adjustment is less sensitive in the initial stage and more flexible in the later stages.

[0144] 5. From Orientation Expand and generate new nodes :

[0145] .

[0146] if If it is a valid node (no collision), add it to the tree and connect and .

[0147] In the current iteration step, the current path tree is generated through the above process. Iterate the above process of generating the path tree until both the starting point and the target point are included in the path tree, stop the iteration, and find the first initial path from the starting point to the target point .

[0148] After finding the first initial path from the starting point to the target point After that, the ellipse / ellipsoid sampling domain is restricted, and finally a better global path is generated.

[0149] 3. Initial Path Optimization

[0150] In some embodiments:

[0151] The Informed strategy is used to optimize the generated initial path to obtain the global path of the robot, which is as follows:

[0152] Calculate the length of the initial path ;

[0153] Use the length of the preliminary path Define an ellipse sampling space, which contains all possible improvement paths;

[0154] Uniformly sample a point within the ellipse , iterate the above RRT* algorithm, and before reaching the set maximum number of iterations, repeatedly update the size of the ellipse sampling space so that the ellipse continues to shrink to achieve the purpose of rapid path convergence.

[0155] Next, the process of Informed-global path optimization in this embodiment will be described in detail:

[0156] 1. Record the initial path Length .

[0157] 2. Define the elliptical sampling space:

[0158] Use the length of the preliminary path To define an elliptical sampling space, the ellipse contains all possible improved paths. The foci of the ellipse are the starting point and end point The length of the major axis of the ellipse is .

[0159] Calculate the distance between the start and end points:

[0160] .

[0161] Length of the major axis of the ellipse:

[0162] .

[0163] Length of the minor axis of the ellipse:

[0164] .

[0165] Center of the ellipse:

[0166] .

[0167] in, , Indicates the horizontal and vertical coordinates of the starting point; , Indicates the horizontal and vertical coordinates of the end point.

[0168] 3. Sampling stage:

[0169] Uniformly sample a point within the ellipse :

[0170] .

[0171] in, is a uniformly distributed random angle , , is a uniformly distributed random number, .

[0172] In the tree Find the The nearest node :

[0173] .

[0174] Generate new nodes according to the previous dynamic step adjustment formula And add it to the tree, the new node The calculation formula is:

[0175] .

[0176] Before reaching the maximum number of iterations, the size of the ellipse sampling space is repeatedly updated so that the ellipse continues to shrink to achieve rapid path convergence.

[0177] 4. Experimental part

[0178] like Figure 2 and Figure 3 As shown, Figure 2 The global path map of the robot generated by the traditional Informed-RRT* method; Figure 3 This is a global path map of the robot generated using the method of the first embodiment.

[0179] exist Figure 2 and Figure 3 middle:

[0180] Dark grey circles and rectangles represent static obstacles;

[0181] The light grey lines represent the exploration paths grown from the RRT;

[0182] The orange dotted ellipse indicates the generated informed ellipse division area;

[0183] The small red dot indicates the starting point;

[0184] The blue dot indicates the endpoint;

[0185] The large red dots represent obstacles that grow near static obstacles;

[0186] The green path indicates the final generated path;

[0187] Figure 2 and Figure 3 The cost in represents the time consumed.

[0188] By comparison Figure 2 and Figure 3 , we can know that:

[0189] Figure 2 Generating the initial path does not produce a goal-oriented effect, causing the growth tree to grow randomly and consume more time (using more nodes), so that the informed optimization strategy cannot be used to further optimize the path length.

[0190] Figure 3It has purposeful growth, and the initial path takes less time (fewer nodes are used), so there is more time (nodes) for informed strategy optimization. So as can be seen from the figure, Figure 3 The path length is significantly better than Figure 2 .

[0191] Table 1 shows the data of 10 times global path planning using the improved informed-RRT* of this embodiment.

[0192] Table 1: Data of 10 global path planning after improved informed-RRT*

[0193]

[0194] Table 2 shows the data of the original informed-RRT* 10-time global path planning.

[0195] Table 2: Data of 10 global path planning of the original informed-RRT*

[0196]

[0197] From the comparison between Table 1 and Table 2, it can be seen that the improved informed-RRT* has a smaller average path loss, and the planning success rate is higher and the time consumption is shorter under the condition of a limited number of nodes, so the scheme has better practicality.

[0198] Embodiment 2

[0199] This embodiment discloses a robot global path planning system based on probability interval division.

[0200] The robot global path planning system based on probability interval division includes:

[0201] The area statistics module is configured to: obtain the total map area of ​​the robot's global path planning, divide the total map area into N equal parts, and count the obstacle area in each equal division and the total obstacle area;

[0202] The first sampling probability calculation module is configured to: calculate the sampling probability within the obstacle-based interval based on the obstacle area and the total obstacle area within each equally divided interval;

[0203] The second sampling probability calculation module is configured to: determine the positions of the starting point and the end point, and the equally divided interval including the starting point and the end point, and set the sampling probability within the interval based on the position of the equally divided interval including the starting point and the end point;

[0204] The total sampling probability calculation module is configured to: combine the sampling probability within the interval based on obstacles and the sampling probability within the interval based on positions to obtain the sampling probability within the area;

[0205] The initial path generation module is configured as follows: based on the RRT* algorithm, the sampling probability in the region is applied to select nodes, and the node closest to the selected node is found, and the equally divided interval is sampled in a variable step size manner to obtain a new node extending from the node closest to the selected node to the selected node, and the new node is connected to the node closest to the selected node to obtain the current path tree;

[0206] The iteration module is configured to: iterate the initial path generation module until the starting point and the target point are both included in the path tree, stop the iteration, and obtain the initial path;

[0207] The path optimization module is configured to optimize the generated initial path to obtain the global path of the robot.

[0208] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0209] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A robot global path planning method based on probability interval division, characterized in that: The following steps are involved: Step 1: Obtain the total map area of ​​the robot's global path planning, divide the total map area into N equal parts, and count the obstacle area in each divided interval and the total obstacle area; Step 2: Based on the obstacle area in each equally divided interval and the total obstacle area, calculate the sampling probability in the interval based on the obstacle; Step 3: Determine the positions of the starting point and the end point, and the equally divided intervals including the starting point and the end point, and set the sampling probability within the interval based on the position of the equally divided intervals including the starting point and the end point; Step 4: Combine the sampling probability within the interval based on obstacles and the sampling probability within the interval based on positions to obtain the sampling probability within the area, specifically: ； in, For the The sampling probability within the region between the equal partitions; For the The probability of sampling within the obstacle-based interval between equal partitions; For the The sampling probability of the interval based on the position between the equal partitions; and is the weight; set up: ； ； but: ； Step 5: Based on the RRT* algorithm, apply the sampling probability in the area to select nodes, and find the node closest to the selected node. Use the variable step size method to sample the equally divided intervals to obtain new nodes extending from the node closest to the selected node to the selected node, and connect the new node and the node closest to the selected node to obtain the current path tree; Step 6: Iterate the above step 5 until both the starting point and the target point are included in the path tree, stop the iteration, and obtain the initial path; Step 7: Optimize the generated initial path to obtain the global path of the robot.

2. The robot global path planning method based on probability interval division as claimed in claim 1, characterized in that: In step 2, based on the obstacle area in each equally divided interval and the total obstacle area, the sampling probability in the obstacle-based interval is calculated. The specific formula is: ； in, is the sampling probability within the interval based on obstacles, Indicates Equally divided intervals; For the The obstacle area between the equal partitions; is the total number of equally divided intervals; The cumulative sum is Obstacle area.

3. The robot global path planning method based on probability interval division as claimed in claim 1, characterized in that: In step 3, the sampling probability within the interval based on the position is set between the equal partitions including the starting point and the end point, including: The sampling probabilities based on the position in the intervals between the equal partitions including the starting point and the end point are all set to 0.5, and the sampling probabilities based on the position in the intervals between the other equal partitions are all set to 0.

4. The robot global path planning method based on probability interval division as claimed in claim 1, characterized in that: The step five specifically includes: Setting up RRT Multiple parameters of the algorithm, and randomly select a node in the map using the sampling probability in the obtained area ; Find the distance in the initial tree The nearest node ; calculate and Distance , using a greedy step size adjustment strategy based on distance Adjust the step length ; from Orientation Expand and generate new nodes ; judge Is it a valid node without collision? If so, Add to the tree and connect and ; Cycle the above process to find the first initial path from the starting point to the target point .

5. The robot global path planning method based on probability interval division as claimed in claim 4, characterized in that: calculate and Distance , using a greedy step size adjustment strategy based on distance Adjust the step length , the specific formula is: ； ； ； ； in, is the adjusted step length of each step; is the initial step length; and Parameters adjusted to control step size; and for The minimum and maximum values ​​of and for The minimum and maximum values ​​of is the current iteration number; is the maximum number of iterations; is the average length of the paths in the current tree; is the initial average path length; is the expected path length to the target area; is the target area.

6. The robot global path planning method based on probability interval division as claimed in claim 4, characterized in that: from Orientation Expand and generate new nodes , the specific formula is: 。 7. The robot global path planning method based on probability interval division as claimed in claim 4, characterized in that: In the map, randomly select a node using the sampling probability within the region , specifically: Generate a random number; Based on the sampling probability within each equal partition, the cumulative sampling probability between the equal partitions is calculated; Determine the cumulative sampling probability of the specific equal-division interval into which the generated random number falls, and select the equal-division interval based on this; Generate random points in the selected equally divided interval as nodes .

8. The robot global path planning method based on probability interval division as claimed in claim 1, characterized in that: The Informed strategy is used to optimize the generated initial path to obtain the global path of the robot.

9. A robot global path planning system based on probability interval division, characterized by: include: The area statistics module is configured to: obtain the total map area of ​​the robot's global path planning, divide the total map area into N equal parts, and count the obstacle area in each equal division and the total obstacle area; The first sampling probability calculation module is configured to: calculate the sampling probability within the obstacle-based interval based on the obstacle area and the total obstacle area within each equally divided interval; The second sampling probability calculation module is configured to: determine the positions of the starting point and the end point, and the equally divided interval including the starting point and the end point, and set the sampling probability within the interval based on the position of the equally divided interval including the starting point and the end point; The total sampling probability calculation module is configured to combine the sampling probability within the interval based on obstacles and the sampling probability within the interval based on positions to obtain the sampling probability within the area, specifically: ； in, For the The sampling probability within the region between the equal partitions; For the The probability of sampling within the obstacle-based interval between equal partitions; For the The sampling probability of the interval based on the position between the equal partitions; and is the weight; set up: ； ； but: ； The initial path generation module is configured as follows: The algorithm selects nodes by using sampling probability in the area, and finds the node closest to the selected node. It samples the equally divided intervals by using variable step size to obtain new nodes extending from the node closest to the selected node to the selected node, and connects the new node with the node closest to the selected node to obtain the current path tree. The iteration module is configured to: iterate the initial path generation module until the starting point and the target point are both included in the path tree, stop the iteration, and obtain the initial path; The path optimization module is configured to optimize the generated initial path to obtain the global path of the robot.

Images