Unmanned ship path planning method based on improved RRT and artificial potential field method

Abstract

The invention discloses an unmanned ship path planning method based on an improved RRT and an artificial potential field method, and the method comprises the steps: obtaining the navigation information of an unmanned ship, and determining a starting point and a target point of the unmanned ship; based on an improved RRT algorithm, two-way extension trees are alternately constructed, wherein the two-way extension trees comprise a starting point extension tree and a target point extension tree; when the starting point extension tree and the target point extension tree meet the meeting condition, backtracking from the meeting node to generate a global path from the starting point to the target point; and smoothing the generated global path, and outputting an optimal path. According to the method, the randomness of the RRT algorithm and the optimization capability of the APF are combined, two-way extension tree construction and a target deviation sampling mechanism are utilized, optimization of path planning of the unmanned ship in the complex water area is achieved, in the path planning process, the global optimal path can be quickly searched, the smooth local path can be generated, and the path planning efficiency is improved. And the adaptability of the unmanned surface vehicle to obstacles is enhanced by introducing a dynamic obstacle avoidance strategy.

Description

Technical Field

[0001] The invention belongs to the technical field of intelligent navigation and path planning, relates to unmanned boat path planning, and specifically relates to an unmanned boat path planning method based on improved RRT and artificial potential field method. Background Art

[0002] With the widespread application of unmanned boats, they have played an important role in the fields of marine exploration, environmental monitoring and rescue. Unmanned boat path planning is a key technology for autonomous navigation of unmanned boats. Its purpose is to plan a safe, efficient and smooth path from the starting point to the target point for the unmanned boat, and at the same time, it is required to be able to bypass obstacles encountered during navigation. Existing path planning algorithms mainly include the rapidly expanding random tree (RRT) algorithm and the artificial potential field method (APF). Among them, the RRT algorithm has good randomness and global search capabilities, and can quickly generate feasible paths, but the paths generated by the algorithm are usually too long and not smooth, especially in complex environments, which is prone to redundant nodes, reducing the efficiency of path planning and the navigation quality of the unmanned boat; while the artificial potential field method guides path generation by constructing target gravity and obstacle repulsion fields. The path smoothness is good, but it is easy to fall into the local optimal solution and has poor adaptability to dynamic obstacles. In practical applications, a single algorithm is often difficult to simultaneously meet the requirements of path length, planning time, obstacle avoidance ability and path smoothness.

[0003] Therefore, a new technical solution is needed to solve these problems. Summary of the invention

[0004] Purpose of the invention: In order to overcome the shortcomings of the prior art, a path planning method for an unmanned boat based on an improved RRT and artificial potential field method is provided, which combines the randomness of the RRT algorithm and the optimization capability of the APF, and utilizes a bidirectional expansion tree construction and a target bias sampling mechanism to optimize the path planning of the unmanned boat in complex waters. During the path planning process, the method of the invention can not only quickly search for the global optimal path, but also generate a smooth local path, and enhances the adaptability of the unmanned boat to obstacles by introducing a dynamic obstacle avoidance strategy.

[0005] Technical solution: To achieve the above purpose, the present invention provides a path planning method for an unmanned boat based on improved RRT and artificial potential field method, comprising the following steps:

[0006] S1: Obtain the navigation information of the unmanned boat and determine the starting point and target point of the unmanned boat;

[0007] S2: Based on the improved RRT algorithm, bidirectional expansion trees are constructed alternately, including the starting point expansion tree and the target point expansion tree;

[0008] S3: When the starting point expansion tree and the target point expansion tree meet the meeting condition, backtrack from the meeting node to generate a global path from the starting point to the target point;

[0009] S4: Smooth the generated global path and output the optimal path.

[0010] Furthermore, the navigation information of the unmanned boat in step S1 includes the coordinate positions of the starting point and the target point, the obstacle distribution in the water environment, and the boundary constraints of the navigation area.

[0011] Furthermore, in step S2, alternately constructing a bidirectional expansion tree based on the improved RRT algorithm specifically includes:

[0012] A1: Generate random sampling points X from the starting point S and the target point G at the same time rand , the nodes of the expanded tree gradually move closer to the middle until the two trees meet;

[0013] A2: During the expansion process, select a random sampling point X in the current expansion tree. rand The nearest node X nearest , and based on this, generate a new node X new ;

[0014] A3: For the new node X new Perform feasibility check on the path and add the new node X that passes the check new Added extension tree.

[0015] Furthermore, the improved RRT algorithm in step A1 adopts a target biased sampling mechanism to generate random sampling points X by the following formula: rand :

[0016]

[0017] Among them, X near is the nearest node, X free is a random sampling point in the global free space, P is a random number, and λ is the probability factor of the target deviation.

[0018] Furthermore, in step A2, the new node position X is calculated based on the artificial potential field method. new , and its calculation formula is:

[0019]

[0020] Among them, ξ is the step size parameter, F rep is the obstacle repulsion, and α is the adjustment coefficient.

[0021] Furthermore, the feasibility check in step A3 specifically includes:

[0022] B1: Obstacle distance detection: Calculate the distance ρ(X,X obs ), if ρ(X,X obs )≤safe distance ρ0, then remove the node and regenerate the sampling point;

[0023] B2: If the situation in step B1 occurs, the path direction is corrected through the repulsion model to prevent the unmanned boat from entering the obstacle area;

[0024] B3: Under the joint action of the gravity model and the repulsion model, the path expansion direction is dynamically adjusted.

[0025] Furthermore, the repulsive force model in step B2 utilizes the repulsive force F rep Correct the path direction, and the calculation formula of repulsion is:

[0026]

[0027] The repulsive force model is obtained by using the repulsive potential field U rep (X) represents the repulsive force of the obstacle on the unmanned boat, and its calculation formula is:

[0028]

[0029] Among them, k rep is the gain coefficient of the repulsive field, ρ(X,X obs ) is the distance from the current node X to the obstacle X obs ρ0 is the Euclidean distance of the obstacle.

[0030] Furthermore, in step B3, the gravitational model is realized by the gravitational potential field U att (X) represents the attraction of the target point to the unmanned boat, and its calculation formula is:

[0031]

[0032] Among them, k att is the proportional coefficient of the gravitational field, which is used to adjust the strength of the attraction of the target point; ρ(X,X goal ) is the current node X to the target point X goal The role of gravity is to guide the expansion tree to approach the target point, thereby reducing the circuitous nature of path planning. By calculating the negative gradient of the gravitational potential field, the specific expression of gravity is obtained:

[0033] F att (X) = k att ρ(X,X goal ) goal

[0034] Among them, n goalIt is a unit direction vector pointing to the target point. The magnitude of gravity decreases as the distance between the unmanned boat and the target point decreases. When the unmanned boat approaches the target point, the effect of gravity gradually decreases to zero, ensuring that the unmanned boat can reach the target point smoothly.

[0035] When ρ(X,X obs )≤ρ0, the repulsive force will be significantly enhanced, forcing the unmanned boat to avoid obstacles; when ρ(X,X obs )>ρ0, the repulsive force disappears, ensuring that the unmanned boat is not disturbed when it is away from obstacles. total , the unmanned boat can achieve dynamic obstacle avoidance and approach the target point steadily, and the comprehensive force F total The calculation formula is:

[0036] F total =F att +F rep

[0037] Furthermore, the meeting condition in step S3 is:

[0038] A global path is generated when two expansion trees meet the following connection conditions:

[0039] The distance between any two nodes is less than the set connection threshold.

[0040] Furthermore, the smoothing process in step S4 includes steering angle verification, specifically:

[0041] Entering the steering angle verification mechanism: the system calculates the new node X new , the nearest node X nearest and its parent node X parent The angle θ between them is used to determine the rationality of the path turning; the calculation formula of the angle θ is:

[0042]

[0043] If the steering angle θ exceeds the set threshold θ max , the node is discarded to avoid sharp turns in the path; if the turning angle is less than the threshold, the system will add the new node to the expansion tree and continue to expand. It can eliminate sharp turns in the path, optimize the continuity of the path, and make the unmanned boat sail more stable.

[0044] Beneficial effects: Compared with the prior art, the present invention solves the adaptability and efficiency problems of the path planning algorithm in complex obstacle scenes by introducing a bidirectional expansion tree construction mechanism, a target bias sampling strategy, and an optimization model of the combined effect of gravity and repulsion. Specifically, it has the following advantages:

[0045] 1. Path optimization: The path generated by the method of the present invention is shorter and smoother, reducing the path length by more than 20% compared with the existing RRT algorithm;

[0046] 2. Improved path planning efficiency: The present invention adopts target biased sampling and bidirectional expansion tree to significantly shorten the path planning time and improve the efficiency by 40%;

[0047] 3. Enhanced obstacle avoidance capability: Combined with the repulsion optimization strategy, it effectively avoids dynamic and static obstacles in complex environments, and the obstacle avoidance capability is significantly enhanced;

[0048] 4. Wide applicability: The method of the present invention can be widely used in unmanned boat path planning tasks in complex water environments, such as rescue operations, environmental monitoring, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a flow chart of the method of the present invention;

[0050] Figure 2 This is the simulation result diagram of a simple obstacle environment;

[0051] Figure 3 This is the simulation result diagram of complex obstacle environment. DETAILED DESCRIPTION

[0052] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0053] like Figure 1 As shown, the present invention provides an unmanned boat path planning method based on improved RRT and artificial potential field method, comprising the following steps:

[0054] S1: Obtain the navigation information of the unmanned boat and determine the starting point and target point of the unmanned boat;

[0055] Environmental modeling and initialization are the basic steps for the implementation of the entire algorithm. Their main purpose is to provide accurate environmental information and initial conditions for the path planning of the unmanned boat. The unmanned boat obtains environmental information of the target waters through sensors, radars or preset electronic maps, including water boundaries, obstacle distribution, navigable areas and non-navigable areas. After preprocessing, this information will be converted into a computable mathematical model by the system so that the path planning algorithm can accurately analyze the water environment.

[0056] The first step in environmental modeling is to define the boundaries of the water area. The role of the boundary is to clarify the range of activities of the unmanned boat and ensure that any nodes generated during the path planning process will not cross the boundaries of the water area. The obstacle information in the water area also needs to be accurately collected, including the number, location, shape and size of the obstacles. Obstacles are usually represented as circles, rectangles or other geometric shapes, which are marked as non-navigable areas in the environmental model. By marking the location of obstacles, the system can dynamically avoid these areas in path planning to ensure the navigation safety of the unmanned boat.

[0057] After obtaining environmental information, the system needs to determine the specific locations of the starting point and the target point. The starting point S is usually determined by the current GPS coordinates of the unmanned boat, while the target point G is specified by the mission requirements. The starting point and the target point are marked as the root nodes of the expansion tree in the environmental model. The path planning algorithm starts from these two nodes and expands outwards until the two expansion trees meet. In order to improve the efficiency and stability of the algorithm, environmental initialization also includes parameter setting and optimization. The system will set key parameters such as the step size ξ of the expansion tree, the target deviation probability λ of the random sampling point, and the safe distance ρ0 of the repulsive field according to the scale of the water area, the complexity of the obstacles, and the performance characteristics of the unmanned boat. For example, when the water area is large or the obstacle distribution is simple, the system will appropriately increase the step size ξ to speed up the expansion speed; in areas with complex obstacles, the step size will automatically decrease to ensure the accuracy and safety of the path.

[0058] Water map information: Load water environment information through navigation radar or map data, including the distribution of static obstacles and the movement trajectory of dynamic obstacles.

[0059] Navigation parameter setting: Initialize path planning parameters, including step size ξ, target deviation coefficient λ, gravity weight coefficient k att , repulsion weight coefficient k rep wait.

[0060] S2: Based on the improved RRT algorithm, bidirectional expansion trees are constructed alternately, including the starting point expansion tree and the target point expansion tree;

[0061] The alternating construction of a bidirectional expansion tree based on the improved RRT algorithm specifically includes:

[0062] A1: Generate random sampling points X from the starting point S and the target point G at the same time rand , the nodes of the expanded tree gradually move closer to the middle until the two trees meet;

[0063] The improved RRT algorithm adopts the target biased sampling mechanism and generates random sampling points X through the following formula: rand :

[0064]

[0065] Among them, X near is the nearest node, X free is a random sampling point in the global free space, P is a random number, and λ is the probability factor of the target deviation, where λ is adjusted according to the requirements of the path planning task. Start with a smaller value and gradually increase it, observe the changes in the path planning results, and finally determine a suitable value.

[0066] The path expansion process starts from the starting point and the target point at the same time, and continues to advance through the construction of a bidirectional expansion tree. Each time the path is expanded, the system generates a random sampling point X in the water environment. rand The generation of sampling points is not only random, but also combined with a target biased sampling mechanism. The target biased mechanism improves the efficiency of the expansion tree approaching the target point by setting the probability of the sampling point being close to the target point, while avoiding excessive and ineffective random expansion.

[0067] A2: During the expansion process, select a random sampling point X in the current expansion tree. rand The nearest node X nearest , and based on this, generate a new node X new ;

[0068] Calculate the new node position X based on the artificial potential field method new , and its calculation formula is:

[0069]

[0070] Among them, ξ is the step size parameter, F rep is the obstacle repulsion, α is the adjustment coefficient;

[0071] A3: For the new node X new Perform feasibility check on the path and add the new node X that passes the check new Added extension tree;

[0072] New Node X new The position of depends not only on the distance between the sampling point and the nearest node, but also on the combined effects of gravity and repulsion, as follows:

[0073] The feasibility check specifically includes:

[0074] B1: Obstacle distance detection: Calculate the distance ρ(X,X obs ), if ρ(X,X obs )≤safe distance ρ0, then remove the node and regenerate the sampling point;

[0075] B2: If the situation in step B1 occurs, the path direction is corrected through the repulsion model to prevent the unmanned boat from entering the obstacle area;

[0076] The repulsive force model uses the repulsive force F rep Correct the path direction, and the calculation formula of repulsion is:

[0077]

[0078] The repulsive force model is obtained by using the repulsive potential field U rep (X) represents the repulsive force of the obstacle on the unmanned boat, and its calculation formula is:

[0079]

[0080] Among them, k rep is the gain coefficient of the repulsive field, ρ(X,X obs ) is the distance from the current node X to the obstacle X obs ρ0 is the Euclidean distance of the obstacle.

[0081] B3: Under the combined effect of the gravity model and the repulsion model, the path expansion direction is dynamically adjusted;

[0082] The gravitational model is based on the gravitational potential field U att (X) represents the attraction of the target point to the unmanned boat, and its calculation formula is:

[0083]

[0084] Among them, k att is the proportional coefficient of the gravitational field, which is used to adjust the strength of the attraction of the target point; ρ(X,X goal ) is the current node X to the target point X goal The role of gravity is to guide the expansion tree to approach the target point, thereby reducing the circuitous nature of path planning. By calculating the negative gradient of the gravitational potential field, the specific expression of gravity is obtained:

[0085] F att (X) = k att ρ(X,X goal ) goal

[0086] Among them, n goal It is a unit direction vector pointing to the target point. The magnitude of gravity decreases as the distance between the unmanned boat and the target point decreases. When the unmanned boat approaches the target point, the effect of gravity gradually decreases to zero, ensuring that the unmanned boat can reach the target point smoothly.

[0087] When ρ(X,X obs )≤ρ0, the repulsive force will be significantly enhanced, forcing the unmanned boat to avoid obstacles; when ρ(X,X obs )>ρ0, the repulsive force disappears, ensuring that the unmanned boat is not disturbed when it is away from obstacles.total , the unmanned boat can achieve dynamic obstacle avoidance and approach the target point steadily, and the comprehensive force F total The calculation formula is:

[0088] F total =F att +F rep

[0089] Use the target gravity F att and obstacle repulsion F rep Optimize the expansion direction to make the path smoother.

[0090] During the path extension process, the combined force F total It not only determines the direction of the expansion tree, but also improves the efficiency of path planning by dynamically adjusting the step size ξ of the path expansion. When the unmanned boat approaches an obstacle, the repulsive force increases, prompting the path to be adjusted to a safe area; when the unmanned boat approaches the target point, the gravity guides the expansion tree to converge smoothly to the target point. The present invention achieves efficient path planning and comprehensive improvement of obstacle avoidance capabilities by reasonably designing the strength and range of gravity and repulsive force.

[0091] S3: When the starting point expansion tree and the target point expansion tree meet the meeting condition, backtrack from the meeting node to generate a global path from the starting point to the target point;

[0092] Meeting condition: When two expansion trees meet the following connection conditions, a global path is generated:

[0093] The distance between any two nodes is less than the set connection threshold.

[0094] S4: Smooth the generated global path and output the optimal path.

[0095] Smoothing includes:

[0096] Redundant node elimination: remove unnecessary intermediate nodes in the path;

[0097] Sharp turn adjustment: Optimize the sharp turn points in the path to ensure that the angle does not exceed the maximum allowable value (such as 40°);

[0098] Curve fitting: The cubic spline interpolation method is used to further optimize the path, making the path more continuous and smooth, meeting the actual navigation needs of the unmanned boat.

[0099] Among them, the sharp turn adjustment is achieved through steering angle calibration, specifically:

[0100] Entering the steering angle verification mechanism: the system calculates the new node X new , the nearest node X nearest and its parent node X parentThe angle θ between them is used to determine the rationality of the path turning; the calculation formula of the angle θ is:

[0101]

[0102] If the steering angle θ exceeds the set threshold θ max , the node is discarded to avoid sharp turns in the path; if the turning angle is less than the threshold, the system will add the new node to the expansion tree and continue to expand. It can eliminate sharp turns in the path, optimize the continuity of the path, and make the unmanned boat sail more stable.

[0103] Path optimization and smoothing are key steps in the path planning method of the present invention, which aims to optimize the initial generated path, eliminate redundant nodes, sharp turns and other problems in the path, and thus generate a smooth and efficient final path. The optimized path not only meets the actual navigation needs of the unmanned boat, but also significantly improves its operational stability and safety.

[0104] The optimization of the initial path starts with backtracking the bidirectional expansion tree. When the nodes of the starting point expansion tree and the target point expansion tree meet the connection conditions, the system will start from the connection node and backtrack the starting point tree and the target point tree respectively to generate an initial path. The initial path usually contains many redundant nodes and sharp turns. If these problems are not handled, the unmanned boat path will be long and not smooth, affecting the navigation performance. Therefore, path optimization is a necessary step.

[0105] During the optimization process, the redundant nodes in the path are first detected and eliminated. By determining the linear relationship between consecutive nodes in the path, the intermediate nodes on the straight line of the path are deleted, and only the turning points of the path are retained. This step can effectively reduce the number of nodes in the path and simplify the path structure.

[0106] Then, the sharp turns in the path are adjusted. Sharp turns may cause the unmanned boat to turn unsteadily, or even cause navigation risks. To solve this problem, the system detects the angle between adjacent nodes in the path and compares it with the set maximum angle threshold. If the angle in the path exceeds the threshold, the system introduces an intermediate node in the path segment and reduces the angle by segmented interpolation until it meets the set requirements.

[0107] In order to further improve the continuity and smoothness of the path, the present invention uses polynomial interpolation to smooth the optimized path. During the interpolation process, the system will select key nodes and generate a continuous curve through cubic spline interpolation to make the path smoother and avoid sharp turns that may occur in the path. The final generated path can not only maintain the rationality of the initial path to the greatest extent, but also meet the stability requirements of the unmanned boat in actual navigation.

[0108] In order to verify the effect of the method of the present invention, this embodiment designs a series of simulation experiments for verification, which are as follows:

[0109] The experiment covers two typical water scenarios: simple obstacles and complex obstacles. From the perspective of path planning efficiency and path quality, the experiment compares the APF-Bi-RRT algorithm with the traditional Bi-RRT, APF and A* algorithms, and analyzes their performance in terms of path length, planning time, number of nodes and path smoothness.

[0110] The experimental scene is set in a 1000x1000 two-dimensional water environment. In the simple obstacle scene, the number of obstacles is 3, and in the complex obstacle scene, the number of obstacles increases to 40. The obstacles are distributed with random radii (15-48 units) to simulate the actual obstacle distribution characteristics in the water. The starting point of the unmanned boat is set to (120,85) and the target point is (930.910). Each algorithm runs in different environments and records the key indicators of its path planning.

[0111] like Figure 2 As shown, in a simple obstacle scenario, the APF-Bi-RRT algorithm provided by the present invention significantly optimizes the path length and planning time through target bias sampling and dynamic obstacle avoidance mechanism. The experimental results shown in Table 1 show that the average path length of the APF-Bi-RRT algorithm is 1228.64 meters, the planning time is 3.257 seconds, the number of nodes is 118, and the maximum turning angle is 24°, which are all better than other algorithms. Compared with the Bi-RRT algorithm, the path length is reduced by about 14.5% and the planning time is reduced by about 33%. The traditional Bi-RRT algorithm has low path quality, more sharp turns, and a maximum turning angle of 79°. Although the A* algorithm has a smoother path, the planning time is as long as 6.138 seconds, which is difficult to meet real-time requirements.

[0112] like Figure 3 As shown, in complex obstacle scenarios, the APF-Bi-RRT algorithm provided by the present invention further demonstrates its robustness. By combining the dynamic adjustment mechanism of gravity and repulsion, the expansion tree can quickly avoid obstacle areas and connect target points. The experimental data shown in Table 1 show that in complex obstacle scenarios, the average path length of the APF-Bi-RRT algorithm is 1386.52 meters, the planning time is 4.734 seconds, the number of nodes is 135, and the maximum turning angle is 31°. Compared with the APF algorithm, the path length is reduced by about 8.9% and the planning time is reduced by about 38%. The Bi-RRT algorithm generates a large number of redundant nodes in complex obstacle areas, the planning time is as long as 6.582 seconds, and the path smoothness is poor.

[0113] Table 1

[0114]

[0115]

[0116] Experimental results show that the APF-Bi-RRT algorithm performs well in both simple and complex obstacle scenarios. Its path planning time is significantly lower than that of traditional algorithms, and the number of nodes and path smoothness are also significantly improved. Especially in complex obstacle scenarios, the combination of dynamic obstacle avoidance mechanism and target bias sampling strategy enables the algorithm to quickly adapt to complex environments and generate safe, smooth and efficient paths.

[0117] In dynamic obstacle scenarios, the APF-Bi-RRT algorithm can adjust the repulsive field in real time and effectively avoid interference by dynamically updating the obstacle position. The fluctuations in planning time and path length are controlled within a small range, further verifying its high adaptability and robustness in practical applications.

[0118] It can be seen that the improved algorithm of the present invention has significant advantages in path planning efficiency, path smoothness and calculation time, and is an ideal solution for unmanned boats to achieve efficient autonomous navigation in complex water environments.

Claims

1. A path planning method for an unmanned boat based on improved RRT and artificial potential field method, characterized in that: The steps include: S1: Obtain the navigation information of the unmanned boat and determine the starting point and target point of the unmanned boat; S2: Based on the improved RRT algorithm, bidirectional expansion trees are constructed alternately, including the starting point expansion tree and the target point expansion tree; S3: When the starting point expansion tree and the target point expansion tree meet the meeting condition, backtrack from the meeting node to generate a global path from the starting point to the target point; S4: Smooth the generated global path and output the optimal path.

2. The unmanned boat path planning method based on improved RRT and artificial potential field method according to claim 1 is characterized in that: The navigation information of the unmanned boat in step S1 includes the coordinate positions of the starting point and the target point, the obstacle distribution in the water environment, and the boundary constraints of the navigation area.

3. The unmanned boat path planning method based on improved RRT and artificial potential field method according to claim 1 is characterized in that: The step S2 of alternately constructing a bidirectional expansion tree based on the improved RRT algorithm specifically includes: A1: Generate random sampling points X from the starting point S and the target point G at the same time rand , the nodes of the expanded tree gradually move closer to the middle until the two trees meet; A2: During the expansion process, select a random sampling point X in the current expansion tree. rand The nearest node X nearest , and based on this, generate a new node X new ; A3: For the new node X new Perform feasibility check on the path and add the new node X that passes the check new Added extension tree.

4. The unmanned boat path planning method based on improved RRT and artificial potential field method according to claim 3 is characterized in that: The improved RRT algorithm in step A1 adopts a target biased sampling mechanism to generate random sampling points X by the following formula: rand : Among them, X near is the nearest node, X free is a random sampling point in the global free space, P is a random number, and λ is the probability factor of the target deviation.

5. The unmanned boat path planning method based on improved RRT and artificial potential field method according to claim 3 is characterized in that: In step A2, the new node position X is calculated based on the artificial potential field method. new , and its calculation formula is: Among them, ξ is the step size parameter, F rep is the obstacle repulsion, and α is the adjustment coefficient.

6. The unmanned boat path planning method based on improved RRT and artificial potential field method according to claim 3 is characterized in that: The feasibility check in step A3 specifically includes: B1: Obstacle distance detection: Calculate the distance ρ(X,X obs ), if ρ(X,X obs )≤safe distance ρ0, then remove the node and regenerate the sampling point; B2: If the situation in step B1 occurs, the path direction is corrected through the repulsion model to prevent the unmanned boat from entering the obstacle area; B3: Under the joint action of the gravity model and the repulsion model, the path expansion direction is dynamically adjusted.

7. The unmanned boat path planning method based on improved RRT and artificial potential field method according to claim 6 is characterized in that: In step B2, the repulsive force model uses the repulsive force F rep Correct the path direction, and the calculation formula of repulsion is: The repulsive force model is obtained by using the repulsive potential field U rep (X) represents the repulsive force of the obstacle on the unmanned boat, and its calculation formula is: Among them, k rep is the gain coefficient of the repulsive field, ρ(X,X obs ) is the distance from the current node X to the obstacle X obs ρ0 is the Euclidean distance of the obstacle.

8. The unmanned boat path planning method based on improved RRT and artificial potential field method according to claim 7 is characterized in that: In step B3, the gravitational model is constructed by using the gravitational potential field U att (X) represents the attraction of the target point to the unmanned boat, and its calculation formula is: Among them, k att is the proportional coefficient of the gravitational field, which is used to adjust the strength of the attraction of the target point; ρ(X,X goal ) is the current node X to the target point X goal The Euclidean distance; by calculating the negative gradient of the gravitational potential field, the specific expression of gravity is obtained: F att (X)=k att ρ(X,X goal )n goal Among them, n goal It is a unit direction vector pointing to the target point. The magnitude of gravity decreases as the distance between the unmanned boat and the target point decreases. When the unmanned boat approaches the target point, the effect of gravity gradually decreases to zero, ensuring that the unmanned boat can reach the target point smoothly. The combined force F of attraction and repulsion total , the unmanned boat can achieve dynamic obstacle avoidance and approach the target point steadily, and the comprehensive force F total The calculation formula is: F total =F att +F rep 。 9. The unmanned boat path planning method based on improved RRT and artificial potential field method according to claim 1, characterized in that: The meeting condition in step S3 is: A global path is generated when two expansion trees meet the following connection conditions: The distance between any two nodes is less than the set connection threshold.

10. The unmanned boat path planning method based on improved RRT and artificial potential field method according to claim 1, characterized in that: The smoothing process in step S4 includes steering angle verification, specifically: The system calculates the new node X new , the nearest node X nearest and its parent node X parent The angle θ between them is used to determine the rationality of the path turning; the calculation formula of the angle θ is: If the steering angle θ exceeds the set threshold θ max , the node is discarded to avoid sharp turns in the path; if the turning angle is less than the threshold, the system will add the new node to the expansion tree and continue to expand.

Images