RRTstar algorithm-based path planning method, apparatus and device, and storage medium

Abstract

The invention relates to the field of path planning, in particular to a path planning method and device based on an RRTstar algorithm, equipment and a storage medium. According to the method, path exploration and planning are carried out by adopting an optimized RRT tree from a starting task point and an ending task point at the same time, the path generation speed is increased by means of powerful computing resources of a computer, and the planning efficiency is improved; then the original path depth is smoothed, the problem that the depth fluctuation of a middle path point is too large can be solved, the path is smoother, and for application scenes such as an autonomous underwater robot sensitive to depth, the influence of depth change on the performance and safety of internal instruments of the autonomous underwater robot can be reduced; iteratively optimizing the smooth path to further improve the quality and adaptability of the path; and finally, the navigation state is obtained according to the iterative optimization path, and the final navigation optimization path is obtained through adjustment according to the navigation state, so that the path planning better fits the actual navigation condition, the feasibility and safety of the path are enhanced, and it is guaranteed that the path planning is better from multiple aspects as a whole.

Description

Technical Field

[0001] The present invention relates to the field of path planning, and in particular to a path planning method, device, equipment and storage medium based on an RRT_star algorithm. Background Art

[0002] In recent years, with the development of science and technology, the demand for autonomous underwater robots to operate in the ocean, especially in the open ocean environment, has been growing. In the field of path planning for autonomous underwater robots, traditional planning methods have several technical bottlenecks. First, RRT (Rapidly-exploring Random The path tree generated by the RRT algorithm can quickly explore the environment in path planning, but its tree structure extends too evenly, resulting in a large number of invalid calculations and reducing the efficiency of path planning. In addition, the existing path planning based on the RRT algorithm is often planned from the starting point to the end point point by point. This linear planning method fails to fully utilize the powerful computing resources of modern computers, resulting in low efficiency. Secondly, for the planned path, due to the random expansion characteristics of the RRT tree itself, the depth fluctuation in the middle of the planned path is too large. In fact, most of the time, the underwater robot does not need to frequently change the navigation depth during the sea navigation operation. Frequent changes in navigation depth may affect the performance and safety of the internal instruments of the autonomous underwater robot, which is particularly important in deep-water operations. Furthermore, the current technology fails to fully utilize the idle computing resources of the computer during the navigation of the autonomous underwater robot, resulting in resource waste. At the same time, the branch information in the RRT algorithm path tree except the final path is not effectively utilized. This information actually contains rich environmental characteristics and can assist in adjusting the path. In view of this, developing a new path planning method to solve the above problems and improve the operating efficiency and safety of autonomous underwater robots has important practical significance and broad application prospects. Summary of the invention

[0003] In order to solve the above-mentioned shortcomings in the prior art, the present invention proposes a path planning method, device, equipment and storage medium based on the RRT_star algorithm.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0005] The path planning method based on the RRT_star algorithm includes: using the RRT_star algorithm to create original RRT trees at two adjacent task points of a preset map respectively; optimizing the original RRT tree to obtain an optimized RRT tree; performing bidirectional path planning according to the optimized RRT tree to obtain an original path between the two task points; performing a deep smoothing operation on the original path to obtain a smoothed path; performing an iterative optimization operation on the smoothed path to obtain an iteratively optimized path; obtaining a navigation state according to the iteratively optimized path, and adjusting the iteratively optimized path according to the navigation state to obtain a navigation optimized path. By using the optimized RRT tree, path exploration and planning are carried out simultaneously from the starting task point and the end task point, which makes full use of the powerful computing resources of the computer, further accelerates the path generation speed, and improves the efficiency of path planning; the original path is subjected to a depth smoothing operation to obtain a smoothed path, which can effectively solve the problem of excessive depth fluctuations at the intermediate path points of the planned path, making the path smoother, especially for application scenarios such as autonomous underwater robots that are sensitive to depth changes, and can reduce the impact of frequent depth changes on the performance and safety of the internal instruments of the autonomous underwater robot; then, the smoothed path is iteratively optimized to obtain an iteratively optimized path, which further improves the quality and adaptability of the path; finally, the navigation status is obtained according to the iteratively optimized path, and the iteratively optimized path is adjusted according to the navigation status to obtain the final navigation optimized path. This adjustment makes the path planning more in line with the actual navigation conditions and improves the feasibility and safety of the path.

[0006] Furthermore, the optimization of the original RRT tree to obtain the optimized RRT tree includes: optimizing the step size of the original RRT tree to obtain the optimized primary RRT tree; optimizing the node addition method of the primary RRT tree to obtain the optimized RRT tree. By optimizing the step size of the RRT tree, invalid exploration caused by unreasonable step size can be avoided. The optimized step size can better adapt to different environments and improve the efficiency of path planning; at the same time, the optimized primary RRT tree provides a better basis for subsequent further optimization; optimizing the node addition method of the primary RRT tree can make the tree structure more reasonable, make the tree growth more targeted, further improve the efficiency of path planning, and generate an optimized RRT tree that better meets actual needs.

[0007] Further, the step length of the original RRT tree is optimized to obtain an optimized primary RRT tree, including: randomly obtaining a point on the map to obtain a first random point; judging whether there is an obstacle at the first random point according to the map; when there is no obstacle at the first random point, obtaining the direction of the first random point according to the map to obtain a first direction; obtaining the node closest to the first random point from the original RRT tree to obtain a node P1; obtaining the step length of the original RRT tree to obtain a first step length; taking the node P1 as the starting point, obtaining a point on the map according to the first step length and the first direction to obtain a second random point; judging whether there is an obstacle between the node P1 and the second random point according to the map; if there is no obstacle, returning to execute randomly obtaining a point on the map until a preset first iteration condition is met, then doubling the first step length to obtain the optimized primary RRT tree; if there is an obstacle, returning to execute randomly obtaining a point on the map until a preset second iteration condition is met, then halving the first step length to obtain the optimized primary RRT tree.

[0008] First, a point is randomly generated on the map to obtain the first random point; then, the node closest to the first random point is obtained from the original RRT tree to obtain node P1, which provides a good starting point for subsequent step optimization; and the obstacles between node P1 and the first random point are determined. When there are no obstacles, the operation is repeated until the preset first iteration conditions are met and then the step length is doubled, which can quickly expand the path in the obstacle-free area, avoid invalid calculations caused by uniform extension, and improve planning efficiency; if there are obstacles, the operation is repeated until the preset second iteration conditions are met and then the step length is halved, and detailed exploration is carried out in the obstacle area to reduce invalid attempts. Finally, the optimized primary RRT tree is obtained, which makes the path planning more targeted and efficient, and effectively solves the problem of invalid calculations in the RRT tree.

[0009] Furthermore, the node adding method of the primary RRT tree is optimized to obtain an optimized RRT tree, including: randomly obtaining a point on the map to obtain a third random point; finding a node pnear closest to the third random point from the primary RRT tree; obtaining the direction of the third random point according to the map to obtain a second direction; obtaining its step length from the primary RRT tree to obtain a second step length; taking the node pnear as the starting point, obtaining a node pnew according to the second step length and the second direction; performing node analysis on the primary RRT tree through a node optimization algorithm to obtain a node analysis result, and performing result analysis on the node analysis result; when the node analysis result is that there is a node that meets the preset node conditions If there is no primary RRT tree node that meets the preset node conditions, the primary RRT tree node that meets the preset node conditions is set as the node pmin, and the node pmin is set as the parent node of the node pnew to obtain the optimized RRT tree; when the node analysis result is that there is no primary RRT tree node that meets the preset node conditions, it is determined whether there is an obstacle between the node pnear and the node pnew; if there is no obstacle between the node pnear and the node pnew, the node pnear is set as the parent node of the node pnew to obtain the optimized RRT tree; if there is an obstacle between the node pnear and the node pnew, it returns to execute to randomly obtain a point on the map to obtain the optimized RRT tree. First, the third random point is randomly obtained on the map, and the node pnear closest to the third random point is found from the primary RRT tree. Taking the node as the starting point, a new node is obtained according to the second step length and the second direction, and then analyzed by the node optimization algorithm; if there is a node that meets the preset node conditions, it can be set as the parent node of the new node to obtain an optimized RRT tree, so that the tree structure is more reasonable and the path planning efficiency is improved; if there is no node that meets the conditions and there is no obstacle between the new node and the node pnear, the node pnear is also set as the parent node of the new node to obtain an optimized RRT tree; if there is an obstacle between the node pnear and the node pnew, return to execute to randomly obtain a point on the map to obtain an optimized RRT tree to avoid invalid extension. This series of operations reduces the invalid calculation of the RRT tree, makes path planning more accurate and efficient, and provides a better infrastructure for subsequent path planning, which helps to improve the quality and speed of overall path planning and adapt to complex map environment requirements.

[0010] Furthermore, the bidirectional path planning is performed according to the optimized RRT tree to obtain the original path between the two task points, including: setting any one of the two task points as the first task point; obtaining the direction of the adjacent task point of the first task point based on the first task point and the map to obtain the third direction; randomly generating a first value according to a preset value range, and generating an optimized angle according to the first value and the third direction; obtaining the distance between the first task point and its adjacent task point according to the map to obtain the adjacent distance; obtaining the first range according to the adjacent distance and the preset range adjustment condition; randomly generating a second value according to the first range; obtaining the coordinates of the first task point according to the map; generating optimized coordinates according to the coordinates of the first task point, the optimized angle and the optimized distance value; and generating an optimized angle according to the optimized coordinates. Generate a sampling point on the map, and determine whether the sampling point is in an obstacle; when the sampling point is not in an obstacle, let the optimized RRT tree search whether there is an optimized RRT tree node of an adjacent task point within the preset second range; when the optimized RRT tree does not find the optimized RRT tree node of the adjacent task point, obtain its step length from the optimized RRT tree to obtain a third step length; obtain the direction of the sampling point according to the map; extend the optimized RRT tree to obtain a new node according to the third step length and the direction of the sampling point; when the optimized RRT tree finds the optimized RRT tree node of the adjacent task point, obtain its step length from the optimized RRT tree to obtain the third step length; according to the third step length, extend the optimized RRT tree to intersect with the optimized RRT tree of the adjacent task point to obtain the original path between the two adjacent task points. By selecting any task point as the first task point and obtaining the direction of the adjacent task point, the third direction is determined, providing the initial target orientation for path planning, so that the entire planning process has a clear direction of progress; the first value is randomly generated and combined with the third direction to generate the optimization angle, which ensures the directionality of path planning while maintaining randomness, improves the possibility of expanding the optimized RRT tree to the end point, and makes the path planning more targeted; the adjacent distance between task points is obtained and the first range is obtained according to the preset range adjustment conditions, and the search range is dynamically adjusted to improve the targeting and efficiency of path search; it is judged whether the sampling point is on the obstacle, which improves the safety of path planning and ensures the feasibility of the path; when the sampling point is not on the obstacle, the adjacent task point node is searched by optimizing the RRT tree, which improves the accuracy and efficiency of path planning; if the node is not found, the path is extended to the sampling point direction to build, and if it is found, the original path is intersected, which significantly improves the speed and directness of path planning; these operations make the sampling points more directional and increase the probability of layout towards the end point; the optimized RRT tree can expand towards the end point more efficiently, reduce invalid exploration, improve the accuracy and efficiency of path planning, and provide more reliable path guarantee for achieving task goals.

[0011] Furthermore, the depth smoothing operation is performed on the original path to obtain a smooth path, including: obtaining the depth of the starting task point of the original path according to the map to obtain a first depth; obtaining the depth of the ending task point of the original path according to the map to obtain a second depth; performing depth smoothing on all path points on the original path one by one starting from the starting task point according to the first depth and the second depth to obtain a third depth for each path point; retrieving the path points at the third depth according to the map to obtain a first retrieval result; analyzing the first retrieval result and determining whether there are path points with obstacles in the first retrieval result; when the first retrieval result is that there are no path points with obstacles, generating a smooth path according to the first depth, the second depth and the third depth; otherwise, obtaining the depth of all path points on the original path according to the map to obtain the original depth of each path point; obtaining a fourth depth based on the original depth and the depth of all path points output by a preset depth optimization algorithm; generating a smooth path according to the first depth, the second depth and the fourth depth. First, by obtaining the depth information of the starting and ending task points, a depth reference is provided for the depth balancing operation of the path points; then, depth smoothing is performed on all path points on the original path to ensure that the depth changes of the path points are smooth; by retrieving and analyzing the depth information, it is possible to determine whether there are obstacles at the path points, thereby avoiding potential risks; if there are no obstacles, a smooth path is obtained based on the depth information; if there are obstacles, the depth is re-optimized and a new smooth path is obtained; through a series of depth processing steps, the depth smoothing and obstacle avoidance of the original path are achieved, a balanced path is obtained, and the safety and feasibility of the path are improved; this series of steps not only improves the adaptability and robustness of the path, but also reduces the significant impact of frequent extreme changes in depth on the performance of the autonomous underwater robot, ensuring the safety of the autonomous underwater robot's underwater operations, and has important practical significance for the path planning of the autonomous underwater robot; overall, these steps can optimize the path depth, improve the quality and safety of path planning, and ensure that the path is more in line with actual navigation needs.

[0012] Further, the iterative optimization processing operation on the smooth path to obtain the iterative optimization path includes: obtaining the path length between two task points on the smooth path according to the map to obtain the initial length; using the path planning algorithm to obtain the optimized path and the optimized length corresponding to the optimized path; comparing the initial length and the optimized length, if the optimized length is less than the initial length, then using the optimized length to update the initial length, and returning to execute the path planning algorithm to obtain the optimized path; when the preset iterative stop condition is met, outputting the iterative optimization path corresponding to the optimized length. First, by obtaining the initial length of the path between the two task points, a benchmark is provided for subsequent optimization; then, the path planning algorithm is used to obtain the optimized path and its length, this step is intended to reduce the total length of the path and improve efficiency; by comparing the initial length and the optimized length, if the optimization result is better, then the initial length is updated and the optimization is continued to form an iterative process until the preset iterative stop condition is met; this method not only reduces the path length, improves the economy and efficiency of the path, but also ensures that the final iterative optimization path is output after the conditions are met, and ensures the accuracy and practicality of the path planning; in general, these steps achieve the optimization of the path length and improve the overall performance of the path planning.

[0013] Further, the navigation state is obtained according to the iterative optimization path, and the iterative optimization path is adjusted according to the navigation state to obtain the navigation optimization path, including: obtaining the navigation state according to the iterative optimization path; performing state analysis on the navigation state; when the navigation state is blocked by obstacles, the navigation optimization path is adjusted according to the iterative optimization path. The navigation state is obtained by iteratively optimizing the path, and the state analysis is performed to identify obstacles; once an obstacle is found, the path is adjusted in time to ensure navigation safety and efficiency, thereby optimizing the navigation path and improving the navigation capability of the autonomous underwater robot.

[0014] Furthermore, the path planning device based on the RRT_star algorithm includes:

[0015] The original RRT tree creation module is used to create the original RRT tree at two adjacent task points of the preset map using the RRT_star algorithm; the original RRT tree optimization module is used to optimize the original RRT tree to obtain the optimized RRT tree; the original path planning module is used to perform bidirectional path planning according to the optimized RRT tree to obtain the original path between the two task points; the smooth path generation module is used to perform deep smoothing operations on the original path to obtain a smooth path; the iterative optimization path generation module is used to perform iterative optimization operations on the smooth path to obtain an iterative optimization path; the navigation optimization path generation module is used to obtain the navigation state according to the iterative optimization path, and adjust the iterative optimization path according to the navigation state to obtain the navigation optimization path. The optimized RRT tree is used to realize bidirectional path exploration, which makes full use of computing resources, speeds up path generation, and improves planning efficiency. Deep smoothing effectively reduces path depth fluctuations, which is especially suitable for autonomous underwater robots that are sensitive to depth changes, reducing the impact on instrument performance and safety. Iterative optimization further improves path quality and enhances adaptability. Finally, the path is adjusted according to the navigation state to ensure that the planning fits the actual navigation conditions and improve the feasibility and safety of the path.

[0016] Furthermore, a path planning device based on the RRT_star algorithm comprises: a memory and at least one processor, wherein instructions are stored in the memory; at least one of the processors calls the instructions in the memory so that the path planning device based on the RRT_star algorithm performs each step of the path planning method based on the RRT_star algorithm as described above.

[0017] Furthermore, a computer-readable storage medium stores instructions thereon, wherein when the instructions are executed by a processor, the various steps of the path planning method based on the RRT_star algorithm as described in any one of the above are implemented.

[0018] The beneficial effects of the path planning method based on the RRT_star algorithm of the present invention are:

[0019] The optimized RRT tree is used to simultaneously explore and plan paths from the starting task point and the end task point, which makes full use of the powerful computing resources of the computer, further accelerates the path generation speed, and improves the efficiency of path planning; the original path is subjected to a depth smoothing operation to obtain a smoothed path, which can effectively solve the problem of excessive depth fluctuations at the intermediate path points of the planned path, making the path smoother, especially for application scenarios such as autonomous underwater robots that are sensitive to depth changes, and can reduce the impact of frequent depth changes on the performance and safety of the internal instruments of the autonomous underwater robot; then, the smoothed path is iteratively optimized to obtain an iteratively optimized path, which further improves the quality and adaptability of the path; finally, the navigation status is obtained according to the iteratively optimized path, and the iteratively optimized path is adjusted according to the navigation status to obtain the final navigation optimized path. This adjustment makes the path planning more in line with the actual navigation conditions and improves the feasibility and safety of the path. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A first flow chart of a path planning method based on the RRT_star algorithm provided in an embodiment of the present invention;

[0021] Figure 2 A second flow chart of a path planning method based on the RRT_star algorithm provided in an embodiment of the present invention;

[0022] Figure 3 A third flow chart of a path planning method based on the RRT_star algorithm provided in an embodiment of the present invention;

[0023] Figure 4 A fourth flow chart of a path planning method based on the RRT_star algorithm provided in an embodiment of the present invention;

[0024] Figure 5 A fifth flow chart of a path planning method based on the RRT_star algorithm provided in an embodiment of the present invention;

[0025] Figure 6 A sixth flow chart of a path planning method based on the RRT_star algorithm provided in an embodiment of the present invention;

[0026] Figure 7 A seventh flow chart of a path planning method based on the RRT_star algorithm provided in an embodiment of the present invention;

[0027] Figure 8 An eighth flow chart of a path planning method based on the RRT_star algorithm provided in an embodiment of the present invention;

[0028] Fig. 9A schematic diagram of the structure of a path planning device based on the RRT_star algorithm provided in an embodiment of the present invention;

[0029] Fig.10 A schematic diagram of the structure of a path planning device based on the RRT_star algorithm provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of the path planning method based on the RRT_star algorithm in the embodiment of the present invention includes:

[0033] 101: Use the RRT_star algorithm to create original RRT trees at two adjacent task points on the preset map;

[0034] 102: Optimizing the original RRT tree to obtain an optimized RRT tree;

[0035] 103: Perform bidirectional path planning according to the optimized RRT tree to obtain the original path between the two task points;

[0036] 104: performing a deep smoothing operation on the original path to obtain a smoothed path;

[0037] 105: performing an iterative optimization processing operation on the smooth path to obtain an iterative optimization path;

[0038] 106: Obtaining a navigation state according to the iterative optimization path, and adjusting the iterative optimization path according to the navigation state to obtain a navigation optimization path.

[0039] In this embodiment, the optimization operation of the original RRT tree involves improving and optimizing the step control parameters and node adding method in the ordinary RRT tree; when the actual application scenario encounters the situation where it is necessary to plan the paths between multiple adjacent task points, in order to improve the planning efficiency, the appropriate number of threads should be selected according to the number of CPU cores, and then the multi-threading method is used to perform parallel path planning operations for the paths between these multiple adjacent task points, and the intermediate task point performs a two-way search on the two adjacent task points before and after it, which is equivalent to the two adjacent task points searching and extending toward each other.

[0040] The optimized RRT tree is used to simultaneously explore and plan paths from the starting task point and the end task point, which makes full use of the powerful computing resources of the computer, further accelerates the path generation speed, and improves the efficiency of path planning; the original path is subjected to a depth smoothing operation to obtain a smoothed path, which can effectively solve the problem of excessive depth fluctuations at the intermediate path points of the planned path, making the path smoother, especially for application scenarios such as autonomous underwater robots that are sensitive to depth changes, and can reduce the impact of frequent depth changes on the performance and safety of the internal instruments of the autonomous underwater robot; then, the smoothed path is iteratively optimized to obtain an iteratively optimized path, which further improves the quality and adaptability of the path; finally, the navigation status is obtained according to the iteratively optimized path, and the iteratively optimized path is adjusted according to the navigation status to obtain the final navigation optimized path. This adjustment makes the path planning more in line with the actual navigation conditions and improves the feasibility and safety of the path.

[0041] See also Figure 2 , a second embodiment of the path planning method based on the RRT_star algorithm in the embodiment of the present invention includes:

[0042] 201: Optimizing the step size of the original RRT tree to obtain an optimized primary RRT tree;

[0043] 202: Optimize the node adding method of the primary RRT tree to obtain an optimized RRT tree.

[0044] In this embodiment, by optimizing the step size of the original RRT tree, invalid exploration caused by unreasonable step size can be avoided. The optimized step size can better adapt to different environments and improve the efficiency of path planning. At the same time, the optimized primary RRT tree provides a better basis for subsequent further optimization. By optimizing the node addition method of the primary RRT tree, the tree structure can be more reasonable, the growth of the tree can be more targeted, the path planning efficiency can be further improved, and an optimized RRT tree that better meets actual needs can be generated.

[0045] See also Figure 3 , a third embodiment of the path planning method based on the RRT_star algorithm in the embodiment of the present invention includes:

[0046] 301: randomly obtain a point on the map to obtain a first random point;

[0047] 302: Determine whether there is an obstacle at the first random point according to the map;

[0048] 303: When there is no obstacle at the first random point, obtaining the direction of the first random point according to the map to obtain a first direction;

[0049] 304: Obtain the node closest to the first random point from the original RRT tree to obtain node P1;

[0050] 305: Obtain the step length from the original RRT tree to obtain the first step length;

[0051] 306: Taking the node P1 as the starting point, obtaining a point on the map according to the first step length and the first direction to obtain a second random point;

[0052] 307: Determine whether there is an obstacle between the node P1 and the second random point according to the map;

[0053] 308: If there is no obstacle, return to execute and randomly obtain a point on the map until the preset first iteration condition is met, then double the first step length to obtain an optimized primary RRT tree;

[0054] 309: If there is an obstacle, the process returns to execute to randomly obtain a point on the map until the preset second iteration condition is met, and then the first step length is halved to obtain an optimized primary RRT tree.

[0055] In this implementation, the first iteration condition is to execute step 301 for T0 (preset value) consecutive times; the second iteration condition is to execute step 301 for T1 (preset value) consecutive times, wherein the step size of the RRT tree has an update upper and lower limit, and the preset value in the step size update condition of each RRT tree is different;

[0056] First, in step 301, a point is randomly generated on the map to obtain the first random point; then, in step 304, the node closest to the first random point is obtained from the original RRT tree to obtain node P1, which provides a good starting point for subsequent step optimization; and the obstacle situation between node P1 and the first random point is determined. When there is no obstacle, the operation is repeated until the preset first iteration condition is met and the step length is doubled, which can quickly expand the path in the obstacle-free area, avoid invalid calculations caused by uniform extension, and improve planning efficiency; if there is an obstacle, the operation is repeated until the preset second iteration condition is met and the step length is halved, and detailed exploration is carried out in the obstacle area to reduce invalid attempts. Finally, the optimized primary RRT tree is obtained, which makes the path planning more targeted and efficient, and effectively solves the problem of invalid calculations in the RRT tree.

[0057] See also Figure 4 , a fourth embodiment of the path planning method based on the RRT_star algorithm in the embodiment of the present invention includes:

[0058] 401: Get a random point on the map to get the third random point;

[0059] 402: Find the node pnear closest to the third random point from the primary RRT tree;

[0060] 403: Obtain the direction of the third random point according to the map to obtain a second direction;

[0061] 404: Obtain the step length from the primary RRT tree to obtain a second step length;

[0062] 405: Taking the node pnear as the starting point, obtaining the node pnew according to the second step length and the second direction;

[0063] 406: Performing node analysis on the primary RRT tree by using a node optimization algorithm to obtain a node analysis result, and performing result analysis on the node analysis result;

[0064] 407: When the node analysis result shows that there is a primary RRT tree node that meets the preset node condition, the primary RRT tree node that meets the preset node condition is set as node pmin, and node pmin is set as the parent node of node pnew to obtain an optimized RRT tree;

[0065] 408: When the node analysis result is that there is no primary RRT tree node that meets the preset node condition, determine whether there is an obstacle between the node pnear and the node pnew;

[0066] 409: If there is no obstacle between the node pnear and the node pnew, the node pnear is set as the parent node of the node pnew to obtain an optimized RRT tree;

[0067] 410: If there is an obstacle between the node pnear and the node pnew, return to execute to randomly obtain a point on the map to obtain the optimized RRT tree;

[0068] In this embodiment, the existing node adding method of the original RRT tree (existing ordinary RRT tree) is optimized. The existing node adding method includes the following steps 1, 2, 3, and 4, wherein steps 1, 2, 3, 4, and 5 are steps of the existing algorithm:

[0069] 1. Randomly point Prand in the map;

[0070] 2. Find the node Pnear closest to the random point Prand in the tree T, and extend the distance d from Pnear to the random point Prand to get the node Pnew;

[0071] 3. Test whether there is an obstacle collision between Pnear and Pnew. If yes, return to step 1. If no, set the parent node of Pnew to Pnear.

[0072] 4. Find the node Pmin near Pnew in the tree T (with a distance radius less than r) so that the total length from Pmin to Pnew is the shortest and there is no obstacle collision. If the total length from Pmin to Pnew is shorter than the total length from Pnear to Pnew, change the parent node of node Pnew to node Pmin;

[0073] 5. Find all nodes p1 near Pnew in the tree T. If the total length from Pnew to p1 is shorter than the total length from p1's current parent node to p1 and there is no obstacle collision, change p1's parent node to Pnew.

[0074] Assuming that there are nodes a and b on the primary RRT tree, the total length from node a to node b is the path distance from the starting task point to node a plus the distance from node a to node b; the specific steps of performing node analysis on the primary RRT tree by the node optimization algorithm in step 404 include:

[0075] Determine whether there is a total length from a node to a node pnew that is shorter than a total length from a node pnear to a node pnew within a preset third range of the primary RRT tree based on the node optimization algorithm, so as to obtain a first determination result;

[0076] When there is a total length from a node to the node pnew in the primary RRT tree within the third range that is shorter than the total length from the node pnear to the node pnew; then based on the node optimization algorithm, determine whether there is an obstacle between the node and the node pnew to obtain a second determination result;

[0077] When there is no obstacle between the node and the node pnew, a node analysis result is generated based on the node optimization algorithm, the first determination result, and the second determination result;

[0078] When there is an obstacle between the node and the node pnew, a node analysis result is generated based on the node optimization algorithm, the first determination result, and the second determination result;

[0079] When the primary RRT tree does not have a total length from a node to the node pnew shorter than the total length from the node pnear to the node pnew within the third range, a node analysis result is generated based on the node optimization algorithm and the first determination result;

[0080] First, the third random point is randomly obtained on the map, and the node pnear closest to the third random point is found in the primary RRT tree. Taking the node as the starting point, a new node is obtained according to the second step length and the second direction, and then analyzed by the node optimization algorithm; if there is a node that meets the preset node conditions, it can be set as the parent node of the new node to obtain an optimized RRT tree, so that the tree structure is more reasonable and the path planning efficiency is improved; if there is no node that meets the conditions and there is no obstacle between the new node and the node pnear, the node pnear is also set as the parent node of the new node to obtain an optimized RRT tree; if there is an obstacle between the node pnear and the node pnew, return to execute to randomly obtain a point on the map to avoid invalid extension. This series of operations reduces the invalid calculation of the RRT tree, makes the path planning more accurate and efficient, and provides a better infrastructure for subsequent path planning, which helps to improve the quality and speed of the overall path planning and adapt to the needs of complex map environments.

[0081] See also Figure 5 , a fifth embodiment of the path planning method based on the RRT_star algorithm in the embodiment of the present invention includes:

[0082] 501: Set any one of the two task points as the first task point;

[0083] 502: Acquire directions of adjacent task points of the first task point based on the first task point and the map to obtain a third direction;

[0084] 503: randomly generating a first value according to a preset value range, and generating an optimized angle according to the first value and the third direction;

[0085] 504: Obtain the distance between the first task point and its adjacent task points according to the map to obtain the adjacent distance;

[0086] 505: Obtain a first range according to the adjacent distance and a preset range adjustment condition;

[0087] 506: randomly generate a second value according to the first range;

[0088] 507: Obtain the coordinates of the first task point according to the map;

[0089] 508: Generate optimized coordinates according to the coordinates of the first task point, the optimized angle, and the optimized distance value;

[0090] 509: Generate sampling points on the map according to the optimized coordinates, and determine whether the sampling points are located at obstacles;

[0091] 510: When there is no obstacle at the sampling point, let the optimized RRT tree search whether there is an optimized RRT tree node of an adjacent task point within the preset second range;

[0092] 511: when the optimized RRT tree fails to find the optimized RRT tree node of the adjacent task point, obtain its step length from the optimized RRT tree to obtain a third step length;

[0093] 512: Get the direction of the sampling point according to the map;

[0094] 513: Extend the optimized RRT tree to obtain a new node according to the third step length and the sampling point direction;

[0095] 514: When the optimized RRT tree finds the optimized RRT tree node of the adjacent task point, the step length thereof is obtained from the optimized RRT tree to obtain a third step length;

[0096] 515: Extend the optimized RRT tree to intersect with the optimized RRT tree of the adjacent task point according to the third step length, so as to obtain the original path between the two adjacent task points.

[0097] In this embodiment, it is assumed that the robot needs to pass through multiple task points during a navigation operation. For a task point C, its next adjacent task point D is at a 45° direction, and its previous adjacent task point B is at a -90° direction. If the optimized RRT tree on the task point C needs to be able to expand both toward the task point B and toward the task point D, the expansion range of the optimized RRT tree on the task point C needs to fall within the angle range of ±15° in the direction of the task point B and ±15° in the direction of the task point D (the angle can be set to other appropriate values, here 30° to 60° and -105° to -75°), so that there is a greater possibility that the optimized RRT tree on the task point C will expand toward the adjacent task points of the task point C;

[0098] To facilitate understanding of this embodiment, the following is an example description using task point C, task point D, and task point B. First, a number is randomly generated from 1 to 100 to obtain a value a (501: a first value is randomly generated according to a preset value range), and the probability of each value in the interval being randomly selected is equal. In actual operation, a random number can be randomly selected from a range of 1 to 10000 or larger; the random number is mapped into an angle: if the value a is in the range of 1 to 66, it is mapped to 30° to 60° or -105° to -75° (uniformly distributed, for example, if the randomly generated value a is in the range of 1 to 33, the value a is converted into an angle α, If the value a is in the range of 34 to 66, convert the value a into an angle α. ), if the value a is in the range of 67 to 100, then the value a is mapped to other angles (uniformly distributed). (And the optimized angle is generated according to the first value and the third direction;) That is, the probability that the angle mapped by the value a falls into the range of 30° to 60° (±15° in the D direction) or -105° to -75° (±15° in the B direction) is The probability that the angle mapped by the value a falls into the remaining interval is

[0099] The following is an example of expanding and generating the second random value: if the distance between two adjacent task points is L, (here the distance between task point C and task point D or the distance between task point C and task point B is used as an example), L is updated after Ti times of iterative update operations to obtain the final L, (i is the number of iterative updates of L), L increases by 1.5 times after each iterative update, that is, L = L*1.5, and the maximum value of L shall not exceed the length of the path that has been explored between the two adjacent task points (here the length of the path that has been explored between task point C and task point D or the length of the path that has been explored between task point C and task point B is used as an example); if there is no explored path between the two task points, L shall not exceed the maximum navigation distance of the robot, that is, the distance between the starting task point and the end task point of the robot's current operation; according to the above adjustment method (that is, the range adjustment condition referred to in step 503), the final L is obtained; according to the L generation interval (0 to )(the first range of step 503); from (0 to ) randomly selects a value d (the second value in step 504);

[0100] Assume that the coordinates of the task point C are (cx, cy), then the coordinates of the sampling point can be obtained by the following formula: x=cx+d*cosα, y=cy+d*sinα, and the coordinates of the sampling point are (cx+d*cosα, cy+d*sinα).

[0101] By selecting any task point as the first task point and obtaining the direction of the adjacent task point, the third direction is determined, providing the initial target orientation for path planning, so that the entire planning process has a clear direction of progress; the first value is randomly generated and combined with the third direction to generate the optimization angle, which ensures the directionality of path planning while maintaining randomness, improves the possibility of expanding the optimized RRT tree to the end point, and makes the path planning more targeted; the adjacent distance between task points is obtained and the first range is obtained according to the preset range adjustment conditions, and the search range is dynamically adjusted to improve the targeting and efficiency of path search; it is judged whether the sampling point is on the obstacle, which improves the safety of path planning and ensures the feasibility of the path; when the sampling point is not on the obstacle, the adjacent task point node is searched by optimizing the RRT tree, which improves the accuracy and efficiency of path planning; if the node is not found, the path is extended to the sampling point direction to build, and if it is found, the original path is intersected, which significantly improves the speed and directness of path planning; these operations make the sampling points more directional and increase the probability of layout towards the end point; the optimized RRT tree can expand towards the end point more efficiently, reduce invalid exploration, improve the accuracy and efficiency of path planning, and provide more reliable path guarantee for achieving task goals.

[0102] See also Figure 6 The sixth embodiment of the path planning method based on the RRT_star algorithm in the embodiment of the present invention includes:

[0103] 601: Obtain the depth of the starting task point of the original path according to the map to obtain a first depth;

[0104] 602: Obtain the depth of the end task point of the original path according to the map to obtain a second depth;

[0105] 603: Perform depth smoothing processing on all path points on the original path starting from the starting task point one by one according to the first depth and the second depth to obtain a third depth of each path point;

[0106] 604: searching for a path point at a third depth according to the map to obtain a first search result;

[0107] 605: Analyze the first search result and determine whether there is a path point with an obstacle in the first search result;

[0108] 606: When the first search result indicates that there is no path point with an obstacle, generate a smooth path according to the first depth, the second depth, and the third depth;

[0109] 607: On the contrary, the depths of all path points on the original path are obtained according to the map to obtain the original depth of each path point;

[0110] 608: Obtain a fourth depth based on the original depth and the depths of all path points output by a preset depth optimization algorithm;

[0111] 609: Generate a smooth path according to the first depth, the second depth, and the fourth depth.

[0112] In this embodiment, if the depth of the middle path point between two task points is higher / lower than the depth of the two task points, it can be considered to be directly raised / lowered, that is, the depth balance processing operation is performed on the middle path point; for ease of understanding, the following example is used to explain: Assume that the robot needs to pass through multiple task points during a navigation operation. For two adjacent task points A and B, it is assumed that the planned path between task points A and B is needed as the robot's underwater navigation route, where task point A is the starting point on the navigation route, the depth of task point A is 100, and task point B is the end point. The depth is also 100; assuming that there are only two path points p1 and p2 between task point A and task point B, and the distance from task point A to p1 is equal to the distance from p1 to p2 and equal to the distance from p2 to task point B, the robot's navigation order is: A->p1->p2->B; if the depth of p1 is 150 and the depth of p2 is 50, the depth of the navigation route changes from 100->150->50->100, and the depth changes from 100 to 150, then to 50, and finally back to 100. This frequent fluctuation in depth is caused by the random expansion of the optimized RRT tree itself;

[0113] In fact, in most cases, it is not necessary to perform such frequent depth changes during navigation. In this case, the recursive interpolation method can be used to perform smooth depth processing. When obstacles appear at path points during the smooth depth processing, the path points with obstacles are regarded as points whose depths cannot be changed. The recursive interpolation method is continued to be used to perform smooth depth processing on other path points on the navigation route where no obstacles exist.

[0114] According to the depth of task point A and task point B, p1 is first interpolated and smoothed to 100, that is, it becomes 100->100->50->100, and then p2 is smoothed, and the smoothed value is also 100;

[0115] However, assuming that there is an obstacle at a depth of 100 in p2, it must float up to 50 to bypass it (i.e. it cannot be smoothed to 100). In this case, the depth value of p2 is not modified, and then the depths from task point A to p2 and from p2 to task point B are recursively interpolated and smoothed. According to the depths of task points A and p2 (original depths), the depth of p1 is smoothed to 75. There are no other points between p2 and point B. Finally, the depth change of the path between task points A and B is smoothed to 100->75->50->100 (608: the depths of all path points outputted based on the original depth and the preset depth optimization algorithm to obtain the fourth depth;).

[0116] First, by obtaining the depth information of the starting and ending task points, a depth reference is provided for the depth balancing operation of the path points; then, depth smoothing is performed on all path points on the original path to ensure that the depth changes of the path points are smooth; by retrieving and analyzing the depth information, it is possible to determine whether there are obstacles at the path points, thereby avoiding potential risks; if there are no obstacles, a smooth path is obtained based on the depth information; if there are obstacles, the depth is re-optimized and a new smooth path is obtained; through a series of depth processing steps, the depth smoothing and obstacle avoidance of the original path are achieved, a balanced path is obtained, and the safety and feasibility of the path are improved; this series of steps not only improves the adaptability and robustness of the path, but also reduces the significant impact of frequent extreme changes in depth on the performance of the autonomous underwater robot, ensuring the safety of the autonomous underwater robot's underwater operations, and has important practical significance for the path planning of the autonomous underwater robot; overall, these steps can optimize the path depth, improve the quality and safety of path planning, and ensure that the path is more in line with actual navigation needs.

[0117] See also Figure 7 The seventh embodiment of the path planning method based on the RRT_star algorithm in the embodiment of the present invention includes:

[0118] 701: Obtain the path length between two task points on the smooth path according to the map to obtain an initial length;

[0119] 702: Using a path planning algorithm to obtain an optimized path and an optimized length corresponding to the optimized path;

[0120] 703: Compare the initial length and the optimized length. If the optimized length is smaller than the initial length, the optimized length is used to update the initial length, and the process returns to execute the path planning algorithm to obtain the optimized path.

[0121] 704: When a preset iteration stop condition is met, output an iterative optimization path corresponding to the optimization length.

[0122] In this embodiment, the iteration stop condition is: no better path is found for T2 consecutive times (T2 is a preset value) or the path between the two task points needs to be used; the specific steps of planning the optimized path according to the path planning algorithm are:

[0123] Take the two task points as the foci and the initial length as the major axis to construct an ellipse range;

[0124] By optimizing the RRT tree, bidirectional path planning is performed within the ellipse to obtain the optimized path and the optimized length corresponding to the optimized path.

[0125] The ellipse has a property: for a point p that falls on the edge of the ellipse, the sum of the distances between p and the two foci is equal to the length of the major axis of the ellipse. That is, if a point falls outside the ellipse when exploring a new path, the length of the path passing through this point must be longer than the major axis (the length of the original path), and there is no need to continue to perform subsequent calculations on this point.

[0126] To facilitate understanding of the present embodiment, the present embodiment is explained below through an example: Assuming that the robot needs to pass through multiple task points during a navigation operation, during the navigation process, the path between any two adjacent task points can continue to be calculated and optimized: for the two task points, it has been calculated that there is a path that can be reached. Assuming that the length of this path is l, an ellipse is generated with the two task points as the focus and l as the major axis, and then a new two-way path planning is performed within the ellipse (the nodes of the optimized RRT tree are all within the ellipse). If a path with a shorter length of l0 is found, a new ellipse is generated with the two task points as the focus and l0 as the major axis, and the path iterative optimization operation is continued in a loop to make the planned path between the two task points shorter and shorter, until no better path is found for T2 consecutive times (T2 is a preset value) or the robot has reached the starting point of the two task points, that is, it is necessary to start navigating along the optimized path between the two task points, and end the iterative optimization of the path between the two task points.

[0127] First, by obtaining the initial length of the path between two task points, a benchmark is provided for subsequent optimization; then, the path planning algorithm is used to obtain the optimized path and its length. This step aims to reduce the total length of the path and improve efficiency; by comparing the initial length and the optimized length, if the optimization result is better, the initial length is updated and the optimization is continued, forming an iterative process until the preset iteration stop condition is met; this method not only reduces the path length and improves the economy and efficiency of the path, but also ensures that the final iterative optimization path is output after the conditions are met, ensuring the accuracy and practicality of path planning; overall, these steps achieve the optimization of path length and improve the overall performance of path planning.

[0128] See also Figure 8The seventh embodiment of the path planning method based on the RRT_star algorithm in the embodiment of the present invention includes:

[0129] 801: Obtaining the navigation status according to the iterative optimization path;

[0130] 802: Perform status analysis on the navigation status;

[0131] 803: When the navigation state is that there is an obstacle blocking the navigation, the navigation optimization path is obtained according to the iterative optimization path adjustment.

[0132] In this embodiment, the iterative optimization path adjustment method includes removing the edges of the optimized RRT tree where obstacles exist from the optimized RRT tree, and then continuing to perform path search and adjustment operations based on the original path according to the optimized RRT tree, thereby obtaining an adjusted route to obtain a navigation optimization path. By adopting this method, adjustment operations can be performed based on the original path, saving the computing power of the previously calculated part and avoiding repeated calculation of the calculated route; because the optimized RRT tree has been established on the map and the map has been explored to a certain extent, the existing information can be continued in the original optimized RRT tree to continue to extend the exploration toward the target task point. Since it is a two-way search, the target task point also extends an optimized RRT tree, so the branches of the two optimized RRT trees are relatively close. In fact, the two optimized RRT trees can be quickly reconnected to obtain the adjusted path.

[0133] The navigation status is obtained by iteratively optimizing the path, and status analysis is performed to identify obstacles. Once an obstacle is found, the path is adjusted in time to ensure navigation safety and efficiency, thereby optimizing the navigation path and improving the navigation capability of the autonomous underwater robot.

[0134] The above describes the path planning method based on the RRT_star algorithm in the embodiment of the present invention. The following describes the path planning device based on the RRT_star algorithm in the embodiment of the present invention. Fig. 9 In one embodiment of the present invention, a path planning device based on the RRT_star algorithm includes:

[0135] A path planning device based on the RRT_star algorithm includes:

[0136] The original RRT tree creation module 1 is used to create the original RRT tree at two adjacent task points of the preset map by using the RRT_star algorithm;

[0137] The original RRT tree optimization module 2 is used to optimize the original RRT tree to obtain an optimized RRT tree;

[0138] The original path planning module 3 is used to perform bidirectional path planning according to the optimized RRT tree to obtain the original path between two task points;

[0139] The smooth path generation module 4 is used to perform a deep smoothing operation on the original path to obtain a smooth path;

[0140] Iterative optimization path generation module 5, used for performing iterative optimization processing operations on the smooth path to obtain an iterative optimization path;

[0141] The navigation optimization path generation module 6 is used to obtain the navigation status according to the iterative optimization path, and adjust the iterative optimization path according to the navigation status to obtain the navigation optimization path.

[0142] In this embodiment, an optimized RRT tree is used to implement bidirectional path exploration, which makes full use of computing resources, speeds up path generation, and improves planning efficiency. Depth smoothing effectively reduces path depth fluctuations, which is especially suitable for autonomous underwater robots that are sensitive to depth changes, reducing the impact on instrument performance and safety. Iterative optimization further improves path quality and enhances adaptability. Finally, the path is adjusted according to the navigation status to ensure that the planning fits the actual navigation conditions and improve the feasibility and safety of the path.

[0143] Fig.10 6 is a schematic diagram of the structure of a path planning device based on the RRT_star algorithm provided in an embodiment of the present invention. The path planning device 600 based on the RRT_star algorithm may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 613 (for example, one or more processors) and a memory 620, and one or more storage media 630 (for example, one or more mass storage devices) storing application programs 633 or data 632. Among them, the memory 620 and the storage medium 630 can be short-term storage or permanent storage. The program stored in the storage medium 630 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the path planning device 600 based on the RRT_star algorithm. Furthermore, the processor 610 can be configured to communicate with the storage medium 630, and execute a series of instruction operations in the storage medium 630 on the path planning device 600 based on the RRT_star algorithm to implement the steps of the decision-making auxiliary method provided in the above-mentioned method embodiments.

[0144] The path planning device 600 based on the RRT_star algorithm may also include one or more power supplies 640, one or more wired or wireless network interfaces 650, one or more input and output interfaces 660, and / or one or more operating systems 631, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will appreciate that Fig.10 The illustrated structure of the path planning device based on the RRT_star algorithm does not constitute a limitation on the path planning device based on the RRT_star algorithm, and may include more or fewer components than illustrated, or a combination of certain components, or a different arrangement of components.

[0145] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of a path planning method based on the RRT_star algorithm.

[0146] Finally, it should be noted that the above description is only a preferred example of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A path planning method based on the RRT_star algorithm, characterized in that: include: The RRT_star algorithm is used to create the original RRT tree at two adjacent task points on the preset map; Optimizing the original RRT tree to obtain an optimized RRT tree; Perform bidirectional path planning based on the optimized RRT tree to obtain the original path between the two task points; Performing a deep smoothing operation on the original path to obtain a smoothed path; Performing iterative optimization processing operations on the smooth path to obtain an iterative optimization path; The navigation status is obtained according to the iterative optimization path, and the iterative optimization path is adjusted according to the navigation status to obtain the navigation optimization path.

2. The path planning method based on the RRT_star algorithm according to claim 1, characterized in that: The step of optimizing the original RRT tree to obtain an optimized RRT tree includes: Optimizing the step size of the original RRT tree to obtain an optimized primary RRT tree; The node adding method of the primary RRT tree is optimized to obtain an optimized RRT tree.

3. The path planning method based on the RRT_star algorithm as claimed in claim 2, characterized in that: The step size of the original RRT tree is optimized to obtain an optimized primary RRT tree, including: Get a random point on the map to get the first random point; Determine whether there is an obstacle at the first random point according to the map; When there is no obstacle at the first random point, obtaining the direction of the first random point according to the map to obtain a first direction; Obtain the node closest to the first random point from the original RRT tree to obtain node P1; Get the step length from the original RRT tree to get the first step length; Taking node P1 as the starting point, obtain a point on the map according to the first step length and the first direction to obtain the second random point; Determine whether there is an obstacle between node P1 and the second random point according to the map; If there is no obstacle, the process returns to randomly obtain a point on the map until the preset first iteration condition is met, then the first step length is doubled to obtain the optimized primary RRT tree; If there is an obstacle, the process returns to randomly obtain a point on the map until the preset second iteration condition is met, and then the second step length is halved to obtain the optimized primary RRT tree.

4. The path planning method based on the RRT_star algorithm as claimed in claim 2, characterized in that: The method of optimizing the node adding method of the primary RRT tree to obtain an optimized RRT tree includes: Get a random point on the map to get the third random point; Find the node pnear closest to the third random point from the primary RRT tree; Obtaining the direction of the third random point according to the map to obtain the second direction; Get the step length from the primary RRT tree to get the second step length; Take the node pnear as the starting point and obtain the node pnew according to the second step length and the second direction; Perform node analysis on the primary RRT tree through a node optimization algorithm to obtain node analysis results, and perform result analysis on the node analysis results; When the node analysis result shows that there is a primary RRT tree node that meets the preset node condition, the primary RRT tree node that meets the preset node condition is set as node pmin, and node pmin is set as the parent node of node pnew to obtain an optimized RRT tree; When the node analysis result is that there is no primary RRT tree node that meets the preset node conditions, determine whether there is an obstacle between the node pnear and the node pnew; If there is no obstacle between the node pnear and the node pnew, the node pnear is set as the parent node of the node pnew to obtain the optimized RRT tree; If there is an obstacle between the node pnear and the node pnew, the process returns to randomly obtain a point on the map to obtain the optimized RRT tree.

5. The path planning method based on the RRT_star algorithm according to claim 1, characterized in that: The bidirectional path planning is performed according to the optimized RRT tree to obtain the original path between the two task points, including: Set any of the two mission points as the first mission point; Based on the first task point and the map, directions of adjacent task points of the first task point are acquired to obtain a third direction; Randomly generate a first value according to a preset value range, and generate an optimized angle according to the first value and a third direction; Obtain the distance between the first task point and its adjacent task points according to the map to obtain the adjacent distance; Obtaining a first range according to the adjacent distance and a preset range adjustment condition; Randomly generate a second value based on the first range; Get the coordinates of the first mission point according to the map; Generate optimized coordinates according to the coordinates of the first task point, the optimized angle and the optimized distance value; Generate sampling points on the map according to the optimized coordinates, and determine whether the sampling points are in the obstacles; When there is no obstacle at the sampling point, let the optimized RRT tree search whether there is an optimized RRT tree node of an adjacent task point within the preset second range; When the optimized RRT tree fails to find the optimized RRT tree node of the adjacent task point, the step length is obtained from the optimized RRT tree to obtain the third step length; Get the direction of the sampling point according to the map; According to the third step length and the direction of the sampling point, the optimized RRT tree is extended to obtain a new node; When the optimized RRT tree finds the optimized RRT tree node of the adjacent task point, its step length is obtained from the optimized RRT tree to obtain the third step length; According to the third step length, the optimized RRT tree is extended to intersect with the optimized RRT tree of the adjacent task point to obtain the original path between the two adjacent task points.

6. The path planning method based on the RRT_star algorithm according to claim 1, characterized in that: The deep smoothing operation is performed on the original path to obtain a smoothed path, comprising: Obtaining the depth of the starting task point of the original path according to the map to obtain a first depth; Obtain the depth of the end task point of the original path according to the map to obtain a second depth; Performing depth smoothing processing on all path points on the original path one by one starting from the starting task point according to the first depth and the second depth to obtain a third depth of each path point; Retrieving the path point at the third depth according to the map to obtain a first search result; Analyze the first search result and determine whether there is a path point with an obstacle in the first search result; When the first search result is that there is no path point with an obstacle, a smooth path is generated according to the first depth, the second depth, and the third depth; Conversely, the depth of all path points on the original path is obtained according to the map to obtain the original depth of each path point; The depth of all path points outputted based on the original depth and a preset depth optimization algorithm to obtain a fourth depth; A smooth path is generated based on the first depth, the second depth, and the fourth depth.

7. The path planning method based on the RRT_star algorithm according to claim 1, characterized in that: The iterative optimization processing operation is performed on the smooth path to obtain the iterative optimization path, including: Obtain the path length between two task points on the smooth path according to the map to obtain the initial length; A path planning algorithm is used to obtain an optimized path and an optimized length corresponding to the optimized path; Compare the initial length and the optimized length. If the optimized length is smaller than the initial length, use the optimized length to update the initial length, and return to execute the path planning algorithm to obtain the optimized path. When the preset iteration stop condition is met, the iterative optimization path corresponding to the optimization length is output.

8. The path planning method based on the RRT_star algorithm according to claim 1, characterized in that: The step of acquiring the navigation status according to the iterative optimization path and adjusting the iterative optimization path according to the navigation status to obtain the navigation optimization path includes: Obtain navigation status according to the iterative optimization path; Conduct status analysis on navigation status; When the navigation state is that there is an obstacle blocking the way, the navigation optimization path is obtained according to the iterative optimization path adjustment.

9. A path planning device based on the RRT_star algorithm, characterized in that: include: The original RRT tree creation module is used to create the original RRT tree at two adjacent task points of the preset map using the RRT_star algorithm; The original RRT tree optimization module is used to optimize the original RRT tree to obtain an optimized RRT tree; The original path planning module is used to perform bidirectional path planning based on the optimized RRT tree to obtain the original path between two task points; A smooth path generation module is used to perform a deep smoothing operation on the original path to obtain a smooth path; An iterative optimization path generation module is used to perform iterative optimization processing operations on the smooth path to obtain an iterative optimization path; The navigation optimization path generation module is used to obtain the navigation status according to the iterative optimization path, and adjust the iterative optimization path according to the navigation status to obtain the navigation optimization path.

10. A path planning device based on the RRT_star algorithm, characterized in that: The path planning device based on the RRT_star algorithm includes: a memory and at least one processor, wherein instructions are stored in the memory; At least one of the processors calls the instructions in the memory to enable the path planning device based on the RRT_star algorithm to perform the various steps of the path planning method based on the RRT_star algorithm as described in any one of claims 1-8.

11. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the steps of the path planning method based on the RRT_star algorithm as described in any one of claims 1 to 8 are implemented.

Images