Path planning methods, devices, storage media and equipment for unmanned surface vessels

Abstract

This application discloses a path planning method, apparatus, storage medium, and device for unmanned surface vessels (USVs), relating to the field of transportation and vessel technology. The method includes: establishing an environmental map based on the vessel's current position, destination position, vessel status, and obstacle position information; generating an initial path point sequence based on vessel motion characteristic constraints using an improved asymptotically optimal fast exploration random tree algorithm to eliminate waypoint deviation risks; performing motion constraint-aware path optimization on the initial path point sequence to generate a smooth path that can be efficiently executed by the tracking controller; obtaining multiple sets of optimized path point sequences based on vessel characteristic parameters with different levels of constraints, overcoming the limitation that paths under single motion constraint conditions are difficult to adapt to complex sea conditions; and finally, performing a multi-objective comprehensive evaluation of path quality on the multiple sets of optimized path point sequences to select the optimal target planning path, thus solving the problems of insufficient feasibility and low efficiency in vessel path planning.

Description

Technical Field

[0001] This application relates to the field of transportation and vessel technology, and in particular to path planning methods, devices, storage media and equipment for unmanned vessels. Background Technology

[0002] With the rapid development of global marine resource development and intelligent shipping, unmanned surface vehicles (USVs) have become a research hotspot in the field of intelligent ships due to their wide application in environmental monitoring, military reconnaissance, and maritime search and rescue. Against this backdrop, how to achieve efficient path planning for ships in dynamic obstacle environments has become a focus of attention for both academia and industry.

[0003] Currently, traditional methods are insufficient in adapting to the characteristics of ship motion. Ships are prone to problems such as long voyages, frequent turns, and large turns, which can prevent them from accurately reaching the planned waypoints.

[0004] Therefore, how to improve the feasibility and efficiency of ship route planning under complex sea conditions is an urgent problem to be solved. Summary of the Invention

[0005] The main objective of this application is to provide a path planning method, apparatus, storage medium, and equipment for unmanned vessels, aiming to solve the technical problems of insufficient feasibility and low efficiency of vessel path planning under complex sea conditions.

[0006] To achieve the above objectives, this application proposes a path planning method for unmanned surface vessels, the method comprising:

[0007] An environmental map is created based on the ship's current location, destination location, ship status, and obstacle location information.

[0008] In the environmental map, based on the constraints of ship motion characteristics, an improved asymptotic optimal fast exploration random tree algorithm is used to generate an initial path point sequence.

[0009] The initial path point sequence is subjected to motion constraint-aware path optimization processing to output an optimized path point sequence. The path optimization processing is used to optimize issues such as path redundancy, frequent turns, and direct flight feasibility.

[0010] Multiple sets of optimized path point sequences are obtained based on ship characteristic parameters with different levels of constraints;

[0011] A multi-objective comprehensive evaluation of path quality is performed on multiple sets of optimized path point sequences to select the optimal objective planning path.

[0012] In one embodiment, the step of generating an initial pathpoint sequence in the environmental map based on ship motion characteristic constraints using an improved asymptotically optimal fast exploration random tree algorithm includes:

[0013] A dual-constraint model for ship motion is constructed, which is used to constrain the maximum steering angle and the minimum turning radius;

[0014] Random sampling points are generated using a dynamic hybrid sampling mechanism, which includes path enhancement sampling, target bias sampling, heuristic sampling, and uniform random sampling.

[0015] The random sampling points are expanded to obtain an expanded node set;

[0016] The ship motion dual-constraint model is applied to the extended node set, and collision detection and steering feasibility verification are performed on each node to iteratively generate an initial path point sequence.

[0017] In one embodiment, the step of performing motion constraint-aware path optimization processing on the initial path point sequence and outputting an optimized path point sequence includes:

[0018] The initial optimized path points are obtained by removing intermediate nodes that meet the redundancy condition from the initial path point sequence through path pruning.

[0019] By performing a skip connection on the initial optimized path points, intermediate optimized path points are obtained;

[0020] Based on the redundancy point geometric elimination criterion, the intermediate optimized path points are optimized to obtain an optimized path point sequence.

[0021] In one embodiment, the step of performing a multi-objective comprehensive evaluation of path quality on multiple sets of optimized path point sequences and selecting the optimal objective planning path includes:

[0022] A multi-dimensional evaluation vector is constructed based on path length, average turning angle, and path complexity metrics.

[0023] Based on the effectiveness penalty mechanism and the multidimensional evaluation vector, the comprehensive evaluation value of each group of optimized path point sequences is calculated by a weighted comprehensive evaluation algorithm;

[0024] The optimal target planning path is obtained by filtering based on the comprehensive evaluation value.

[0025] In one embodiment, the step of calculating the comprehensive evaluation value of each group of optimized path point sequences using a weighted comprehensive evaluation algorithm based on the effectiveness penalty mechanism and the multidimensional evaluation vector includes:

[0026] Intermediate evaluation parameters are calculated using an effectiveness penalty mechanism;

[0027] The comprehensive evaluation value of each group of optimized path point sequences is calculated using a weighted comprehensive evaluation algorithm.

[0028] In one embodiment, the step of calculating intermediate evaluation parameters through the validity penalty mechanism includes:

[0029] Define a path identifier vector, which includes valid and invalid states;

[0030] When the optimized path point sequence violates the motion constraints, the corresponding identifier vector is set to an invalid state, and the optimized path point sequence that violates the motion constraints is defined as an invalid path;

[0031] The intermediate evaluation parameters are obtained by multiplying the quality evaluation parameters of invalid paths by the preset penalty coefficients.

[0032] In one embodiment, after the step of performing a multi-objective comprehensive evaluation of path quality on multiple sets of optimized path point sequences and selecting the optimal objective planning path, the method further includes:

[0033] Based on the optimal target planning path, the desired heading angle is calculated using an adaptive guidance strategy;

[0034] The actual heading angle is obtained based on the ship's real-time motion status;

[0035] Based on the deviation between the desired heading angle and the actual heading angle, a fuzzy adaptive PID controller generates a rudder angle command.

[0036] The ship's motion state is updated according to the rudder angle command, and the guidance strategy is executed cyclically for path tracking.

[0037] Furthermore, to achieve the above objectives, this application also proposes a path planning device for an unmanned vessel, the path planning device comprising:

[0038] The environmental map creation module is used to create an environmental map based on the ship's current location information, destination location information, ship status information, and obstacle location information.

[0039] An initial path generation module is used to generate an initial path point sequence in the environmental map based on ship motion characteristic constraints and through an improved asymptotic optimal fast exploration random tree algorithm.

[0040] The optimized path generation module is used to perform motion constraint-aware path optimization processing on the initial path point sequence and output an optimized path point sequence. The path optimization processing is used to optimize issues such as path redundancy, frequent turns, and direct flight feasibility.

[0041] Multiple path generation modules are used to obtain multiple sets of optimized path point sequences based on ship characteristic parameters with different levels of constraints;

[0042] The target path planning module is used to perform a multi-objective comprehensive evaluation of the path quality of multiple sets of optimized path point sequences, and to select the optimal target planning path.

[0043] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the path planning method for unmanned vessels as described above.

[0044] In addition, to achieve the above objectives, this application also proposes a path planning device for an unmanned vessel, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the path planning method for the unmanned vessel as described above.

[0045] One or more technical solutions proposed in this application have at least the following technical effects:

[0046] (1) Based on the ship's current location information, destination location information, ship status information and obstacle location information, establish an environmental map to provide a complete obstacle avoidance basis for path search.

[0047] (2) In the environmental map, based on the constraints of ship motion characteristics, the initial path point sequence is generated by the improved progressive optimal fast exploration random tree algorithm, which solves the problem of missing path curvature constraints caused by the lack of consideration of ship underactuation characteristics in the traditional algorithm. This makes the initial path point sequence strictly conform to the ship's physical maneuvering limit and eliminates the risk that the planned waypoints exceed the ship's motion performance.

[0048] (3) Perform path optimization processing with motion constraint perception on the initial path point sequence and output the optimized path point sequence. This solves the problems of path post-processing disrupting obstacle avoidance safety and curvature discontinuity, and generates a smooth path that can be directly and efficiently executed by the tracking controller.

[0049] (4) Based on the ship characteristic parameters with different constraints, multiple sets of optimized path point sequences are obtained, which breaks through the limitation that the path under a single ship motion constraint is difficult to adapt to complex sea conditions. Finally, the path quality of multiple sets of optimized path point sequences is comprehensively evaluated by multiple objectives, and the optimal target planning path is selected. This avoids the problems of mismatch between the motion performance required to complete path tracking and the ship's working conditions, lengthy paths, and frequent path turning leading to unstable control in unmanned ship path planning. It solves the problems of insufficient feasibility and low efficiency of ship path planning under complex sea conditions. Attached Figure Description

[0050] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0052] Figure 1 This is a flowchart illustrating the first embodiment of the path planning method for the unmanned vessel of this application;

[0053] Figure 2 This is a flowchart illustrating the second embodiment of the path planning method for the unmanned vessel of this application;

[0054] Figure 3 This is a schematic diagram illustrating how waypoints in this embodiment of the application do not consider motion constraint extension;

[0055] Figure 4 A schematic diagram showing the extension of motion constraints for waypoints in the embodiments of this application;

[0056] Figure 5 This is a flowchart illustrating the third embodiment of the path planning method for unmanned vessels in this application;

[0057] Figure 6 This is a schematic diagram of the embodiment of this application before redundancy points are eliminated;

[0058] Figure 7 This is a schematic diagram showing the redundancy points eliminated in the embodiments of this application;

[0059] Figure 8 This is a flowchart illustrating the fourth embodiment of the path planning method for unmanned vessels in this application;

[0060] Figure 9 This is a schematic diagram illustrating the improved RRT-star ship routing results according to an embodiment of this application;

[0061] Figure 10 This is a schematic diagram showing the visualization results of the rotation angle parameters of the ship multi-constraint model in the embodiments of this application;

[0062] Figure 11 This is a schematic diagram showing the visualization results of the ship multi-constraint model index parameters in an embodiment of this application;

[0063] Figure 12 This is a schematic diagram comparing the path planning results of traditional RRT, RRT-Star, and improved RRT-Star in the embodiments of this application;

[0064] Figure 13 This is a flowchart illustrating the ship path planning and tracking control system according to an embodiment of this application;

[0065] Figure 14 This is a schematic diagram of the module structure of the path planning device for an unmanned vessel according to an embodiment of this application;

[0066] Figure 15 This is a schematic diagram of the hardware operating environment involved in the path planning method of the unmanned vessel in the embodiments of this application.

[0067] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0068] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0069] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0070] With the rapid development of global marine resource development and intelligent shipping, unmanned surface vessels (USVs) have become a research hotspot in the field of intelligent ships due to their wide application in environmental monitoring, military reconnaissance, and maritime search and rescue. Real-time path planning and robust tracking control in complex waters are the core challenges restricting the large-scale application of USVs. Against this backdrop, how to achieve efficient path planning and precise tracking control of ships in dynamic obstacle environments has become a focus of attention for both academia and industry.

[0071] Most current path planning algorithms are unsuitable for ships, failing to incorporate improvements based on ship motion characteristics. Since ships need to plan routes as quickly as possible, high computational speed is crucial. However, most current algorithms consume significant time and computational resources when dealing with large search spaces. Traditional improvement methods suffer from two major limitations: first, path planning is inefficient in complex sea conditions, failing to match the ship's physical maneuverability and resulting in insufficient path feasibility; second, post-processing of paths (such as Bézier curve smoothing) may compromise obstacle avoidance safety. Furthermore, most studies have not adequately considered the constraints of USV underactuated characteristics on path curvature, making it difficult for the planned path to be efficiently executed by the tracking controller.

[0072] Therefore, how to improve the feasibility and efficiency of ship route planning under complex sea conditions is an urgent problem to be solved.

[0073] Based on this, embodiments of this application provide a path planning method for unmanned surface vessels, referring to... Figure 1 , Figure 1This is a flowchart illustrating the first embodiment of the path planning method for the unmanned vessel of this application.

[0074] In this embodiment, the path planning method for the unmanned vessel includes steps S10 to S50:

[0075] Step S10: Create an environmental map based on the ship's current location information, destination location information, ship status information, and obstacle location information.

[0076] It should be noted that ship status information can include data such as ship speed (with the motor set to a constant speed) and heading angle, while obstacle location information can include the geometric boundary data and coordinate position information of static obstacles (islands, reefs, and shoals) and dynamic obstacles (other ships). An environmental map can be understood as a dynamic spatial model that integrates the ship's current position, destination, ship status information, and obstacle location information.

[0077] Step S20: In the environmental map, based on the constraints of ship motion characteristics, an initial path point sequence is generated using an improved asymptotic optimal fast exploration random tree algorithm.

[0078] It should be noted that constraints on ship motion characteristics may include the maximum turning angle (the ship's instantaneous turning capability), the minimum turning radius (the curvature limit to avoid loss of control and deviation from the course), and the safety radius (collision buffer distance). The improved asymptotically optimal rapid-exploring random tree algorithm can be understood as a free-space search method based on probabilistic exploration, generating feasible paths guided by constraints. For example, the improved asymptotically optimal rapid-exploring random tree algorithm here can be a multimodal improvement strategy proposed based on the classic rapidly-exploring random tree algorithm (RRT-star) framework. It should be noted that Rapidly-exploring Random Tree (RRT) is a sampling-based path planning algorithm suitable for high-dimensional spaces and unstructured environments. Its core idea is to gradually explore the free space through random sampling and tree structure expansion, ultimately generating a feasible path from the starting point to the destination. RRT-star is an improved version of RRT, proposed to address the limitation of suboptimal planning results. The improved asymptotically optimal fast exploration random tree algorithm in this embodiment is a further improvement on RRT-star. It is designed collaboratively through three levels: kinematic constraint modeling, probabilistic sampling optimization, and path feasibility verification, to address the shortcomings of traditional methods in terms of adaptability to ship motion characteristics. The initial path point sequence can be the original waypoint set output by the improved asymptotically optimal fast exploration random tree algorithm.

[0079] Step S30: Perform motion constraint-aware path optimization processing on the initial path point sequence and output an optimized path point sequence. The path optimization processing is used to optimize issues such as path redundancy, frequent turns, and direct flight feasibility.

[0080] It should be noted that motion-constrained path optimization can be understood as a path smoothing technique based on a ship's maneuverability (turning radius or angle). This can include pruning and geometrically eliminating redundant points in the initial path point sequence, as well as path compression through skip connections. Path optimization can solve problems related to lengthy paths, frequent turns, and the feasibility of straight-line navigation.

[0081] Step S40: Based on ship characteristic parameters with different levels of constraints, multiple sets of optimized path point sequences are obtained.

[0082] It should be noted that ship characteristic parameters with different levels of constraint may include multiple sets of preset maneuverability configurations, and multiple sets of optimized path point sequences are obtained based on the multiple sets of preset maneuverability configurations.

[0083] Step S50: Perform a multi-objective comprehensive evaluation of the path quality of multiple sets of optimized path point sequences, and select the optimal target planning path.

[0084] For example, a multi-objective comprehensive evaluation can be a quantitative assessment system that integrates path length (economy), average turning angle (smoothness), and number of path points (complexity). The optimal objective planning path is the globally optimal path selected from the optimized path point sequence.

[0085] In this embodiment, an environmental map is established based on the ship's current position, destination, ship status, and obstacle locations, providing a complete obstacle avoidance foundation for path search. Within the environmental map, an improved asymptotically optimal fast exploration random tree algorithm is used to generate an initial path point sequence based on ship motion characteristic constraints. This addresses the problem of missing path curvature constraints caused by traditional algorithms failing to consider the ship's underactuated characteristics, ensuring that the initial path point sequence strictly conforms to the ship's physical maneuverability limits and eliminating the risk of planned waypoints exceeding the ship's motion performance. The initial path point sequence undergoes motion constraint-aware path optimization processing, outputting an optimized path point sequence. This solves the problems of post-processing compromising obstacle avoidance safety, path redundancy, frequent turns, and straight-line feasibility, generating a smooth path that can be efficiently executed by the tracking controller. Based on ship characteristic parameters with different levels of constraints, multiple sets of optimized path point sequences are obtained, which overcomes the limitation that paths under single ship motion constraints are difficult to adapt to complex sea conditions. Finally, a multi-objective comprehensive evaluation of path quality is carried out on multiple sets of optimized path point sequences, and the optimal target planning path is selected. This avoids problems such as the mismatch between the motion performance required to complete path tracking and the ship's working conditions, the length of the path, and the instability caused by frequent path turning in unmanned vessel path planning. It solves the problems of insufficient feasibility and low efficiency of ship path planning under complex sea conditions.

[0086] Reference Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the path planning method for unmanned surface vessels in this application, based on the above. Figure 1 The first embodiment shown is followed by a second embodiment of the path planning method for the unmanned vessel of this application.

[0087] In the second embodiment, step S20 includes:

[0088] Step S201: Construct a dual-constraint model for ship motion. The dual-constraint model is used to constrain the maximum steering angle and the minimum turning radius.

[0089] It should be noted that the dual-constraint model of ship motion can be understood as a mechanism for quantifying ship maneuverability. The classic RRT-STAR algorithm neglects the consideration of differential constraints on ship kinematics. This is because if there is a large turning angle between two route segments at a waypoint, it will lead to abrupt turns in the path that the ship cannot execute. If the distance of the route segment formed between any two adjacent waypoints is too short (i.e., less than the turning radius constraint), then when the ship travels at a preset speed, the distance the ship travels from the start of steering will be greater than the path length between waypoints when the bow angle changes to the maximum turning angle. This makes it impossible for the ship to accurately reach the planned waypoint. In other words, if the planned path does not consider the ship's motion characteristics, the unmanned ship will be very likely to overshoot the waypoint and deviate from the route during navigation. Therefore, the dual-constraint model of ship motion in this embodiment ensures that the generated path conforms to the actual motion capability of the ship by setting the maximum turning angle (the limit angle of a single turn) and the minimum turning radius (the minimum radius of curvature required for the ship to turn).

[0090] For example, the dual-constraint model of ship motion can be expressed as formula (1).

[0091] (1)

[0092] In formula (1), This is the turning point between two flight paths. This represents the maximum turning angle of the unmanned vessel. This represents the distance of the path between two waypoints. The minimum turning radius is given by the turning angle. Determined by the relationship between vector spaces, it is expressed as formula (2).

[0093] (2)

[0094] In formula (2), and The unit vectors of adjacent flight segments are represented by formulas (3) and (4), respectively.

[0095] (3)

[0096] (4)

[0097] In formulas (3) and (4), This represents a sequence of nodes along a given continuous path.

[0098] Step S202: Random sampling points are generated using a dynamic hybrid sampling mechanism, which includes path enhancement sampling, target bias sampling, heuristic sampling, and uniform random sampling.

[0099] It should be noted that the dynamic hybrid sampling mechanism primarily addresses the inefficiency of traditional uniform random sampling strategies in constrained spaces. It can be understood as an adaptive random sampling strategy used to improve convergence speed. The dynamic hybrid sampling mechanism can include path-enhancing sampling, target-biased sampling, heuristic sampling, and uniform random sampling.

[0100] For example, the first step is path enhancement sampling. If a path has been found, there is an 80% probability of sampling near the path. The sampling function is expressed as formula (5).

[0101] (5)

[0102] In formula (5), It is a Gaussian distribution, and is Controllable expansion step size, These are random nodes within the existing path. To improve the efficiency of extending towards the target, a target bias sampling is set; when the probability is less than the preset target bias probability... Time (here) (Take 0.5), and the target point can be directly sampled. Heuristic sampling generates biased samples by connecting the starting point to the ending point. Dynamic sampling has a 30% probability of applying heuristic sampling, as expressed in formula (6) and formula (7).

[0103] (6)

[0104] (7)

[0105] In formulas (6) and (7), t For random sampling within the interval (0,1), determine the proportion of positions on the line connecting the start and goal, i.e., when t When =0, Starting from, when t =1 The goal is to bias the sampling points towards the line connecting the starting point and the ending point, thereby accelerating the algorithm's search for the target region. It is a Gaussian distribution, with Centered on the map, random points following a normal distribution are generated around it. Finally, regular random sampling is set to uniformly sample points across the entire map area.

[0106] Step S203: Expand the nodes of the random sampling points to obtain an expanded node set.

[0107] It should be noted that the process of generating new candidate nodes from random sampling points involves connecting the current node with the sampling point and extending it to reachable locations to form an expanded node set. Node expansion is used to expand the path search space.

[0108] For example, such as Figure 3 As shown, since traditional RRT node expansion does not consider motion constraints, this study improves the expansion strategy to conform to the ship's motion characteristics. When expanding the ship's node, the angle difference between the current heading and the target heading needs to be considered. If it exceeds the maximum turning angle, expansion is performed according to the maximum turning angle, such as... Figure 4 As shown.

[0109] First, let the heading angle of the current node be... Waypoints on the target path , They are respectively , Then the expected heading angle It is represented by formula (8).

[0110] (8)

[0111] The formula for calculating and standardizing the relative steering angle is expressed as formula (9).

[0112] (9)

[0113] when At that time, the maximum steering angle is used. The calculation method for limiting the flight path and constraining the heading is expressed as formula (10).

[0114] (10)

[0115] The coordinates of the new node can be determined by the constrained heading, as expressed in formula (11).

[0116] (11)

[0117] In formula (11), Distance The nearest node, For along to A controllable step size for directional expansion. This is the newly generated node.

[0118] Step S204: Apply the ship motion dual-constraint model to the extended node set, perform collision detection and steering feasibility verification on each node, and iteratively generate the initial path point sequence.

[0119] It should be noted that collision detection is used to verify whether a node's position overlaps with an obstacle, and the turning feasibility verification uses a dual-constraint model to check whether the turning angles and radii between adjacent nodes comply with regulations, ensuring a smooth and feasible path. Finally, by repeatedly expanding, verifying, and filtering nodes, an initial path point sequence is gradually constructed.

[0120] For example, the ship motion double constraint model is applied to the extended node set, and the double constraint of the ship motion double constraint model and the collision constraint of the obstacle are realized through discretization, as expressed by formula (12).

[0121] (12)

[0122] In formula (12), and These are functions for steering constraint verification and collision constraint verification, respectively. For nodes in a random tree.

[0123] In this embodiment, a dual-constraint model of ship motion is constructed, with hard constraints on the maximum steering angle and minimum turning radius, deriving a feasible path space that conforms to the ship's kinematic characteristics. This solves the problem of discontinuous path curvature caused by neglecting maneuver constraints in traditional random tree sampling. A dynamic hybrid sampling mechanism is used to dynamically complement local optimization and global guidance, deriving a more efficient sampling point distribution and overcoming the shortcomings of high blindness and slow convergence speed in random sampling. Expanding nodes to generate an extended node set enhances the completeness of the path search. Finally, during the iterative generation process, collision detection and steering feasibility verification are used to perform a two-layer screening of the extended nodes, obtaining an initial path point sequence that considers both ship motion characteristics and convergence speed.

[0124] Reference Figure 5 , Figure 5 This is a flowchart illustrating the third embodiment of the path planning method for unmanned surface vessels in this application, based on the above. Figure 2 The second embodiment shown presents a third embodiment of the path planning method for the unmanned vessel of this application.

[0125] In the third embodiment, step S30 includes:

[0126] Step S301: Remove intermediate nodes that meet the redundancy condition from the initial path point sequence through path pruning to obtain the initial optimized path points.

[0127] It should be noted that path pruning is a redundant node removal strategy. The core idea of ​​path pruning is to remove redundant points in the path and retain key turning points. The criterion for judging whether a point is redundant is: if the path still satisfies the turning constraints and has no collisions after removing the point, then it is considered redundant. By detecting intermediate nodes in the initial path point sequence that meet the redundancy condition (such as collinear nodes or nodes whose change in motion direction is less than a threshold), unnecessary nodes are removed, generating initial optimized path points and compressing the path length.

[0128] For example, the initial path node sequence is first calculated based on the algorithm's node expansion strategy. The path pruning algorithm removes redundant nodes through turning constraint verification and collision constraint verification. Existing candidate nodes are then selected. Its steering angle can be calculated and expressed as formula (13).

[0129] (13)

[0130] In formula (13), and These are represented as formula (14) and formula (15), respectively.

[0131] (14)

[0132] (15)

[0133] It should be noted that if the path is... satisfy ,or If a collision occurs on the path segment, or if the path length is less than twice the minimum turning radius, the current node needs to be retained. Conversely, skipping intermediate nodes is allowed. .

[0134] Step S302: Perform jump connections on the initial optimized path points to obtain intermediate optimized path points.

[0135] It's important to note that skip connections can be considered a path smoothing technique. Its core idea is to attempt to connect distant points while satisfying turning constraints and avoiding collisions. A reverse verification strategy is used to find the maximum feasible span, reducing the number of intermediate points in the path. Essentially, this is a greedy strategy, always attempting to connect the farthest point directly reachable from the current point. Long-distance straight-line connection tests can be performed on the pruned initial optimized path points. If there are no obstacles between two points and the motion constraints are met, intermediate nodes are skipped and the connection is direct, generating intermediate optimized path points and reducing path turns.

[0136] For example, the maximum span can be set to... ,in The maximum jump distance, The number of generated path points is used to reverse the search for the first node that satisfies the ship motion constraints and collision constraints, thereby achieving path compression, as expressed in formula (16).

[0137] (16)

[0138] In formula (16), To optimize the path nodes and enhance the safety of the planned path, we approximate the area of ​​obstacles as a circular region. Then, treating the ship as a point mass, its safety radius is set according to the ship's size. With a safety factor of 1.2, the safe distance for the obstacle's expansion can be obtained as follows: This means that the collision constraint effect is enhanced by increasing the detection distance.

[0139] It should be noted that if the node skipping phase has already processed up to the last point, then add the last point to the optimization path. In the middle, another collision detection is performed to check the connection between it and the endpoint to ensure that it will not collide with obstacles. If a direct connection causes a collision, a midpoint is inserted. The midpoint coordinates are selected as the average of the sum of the coordinates of the last point of the current optimized path and the last point of the original path. If the midpoint does not work, try adding the second to last point of the original path.

[0140] Step S303: Based on the redundancy point geometric elimination criterion, the intermediate optimized path points are optimized to obtain the optimized path point sequence.

[0141] It should be noted that the redundant point geometric elimination criterion targets path points after skip connections. Based on the geometric constraint of minimum turning radius, it deletes or adjusts points that do not conform to the continuity of straight-line navigation of the ship, and generates the final optimized path point sequence.

[0142] For example, nodes are merged when the distance between adjacent nodes is less than 20% of the minimum turning radius, as shown in formula (17).

[0143] (17)

[0144] If the current point is too close to the previous point, skip the current point. It is possible to process three consecutive nodes simultaneously. The direction vector is calculated and expressed as formula (18) and formula (19).

[0145] (18)

[0146] (19)

[0147] It should be noted that when the directional deviation is less than 5°, that is... This means that the three points are approximately collinear, and the intermediate point can be removed. The schematic diagram can be referenced. Figure 6 (Before redundancy elimination) and Figure 7 (After eliminating redundant points). Finally, the obstacle collision constraint function can be used again to check that the entire path will not collide with obstacles. After a series of optimizations, the optimized path node sequence is obtained. .

[0148] In this embodiment, path pruning is used to remove redundant nodes, solving the problem of excessive nodes and low path efficiency caused by sampling redundancy in the initial path point sequence. By deriving node collinearity and convergence of motion direction, invalid intermediate points are directly reduced. By using skip connections to connect optimized points with unobstructed straight lines, the problem of excessive path polylines and abrupt curvature changes is solved. Long-distance straight paths are derived to replace multiple polylines, reducing the number of turning points. Combined with the geometric elimination criterion for redundant points, the geometric compliance of the path after skipping is optimized. By merging nodes with excessively small spacing and deleting approximately collinear intermediate points, the problem of straight-line feasibility caused by node redundancy is solved, ensuring that frequent heading adjustments are reduced and that the ship sails smoothly.

[0149] Reference Figure 8 , Figure 8 This is a flowchart illustrating the fourth embodiment of the path planning method for unmanned surface vessels in this application, based on the above. Figure 5 The third embodiment shown presents a fourth embodiment of the path planning method for the unmanned vessel of this application.

[0150] In the fourth embodiment, step S50 includes:

[0151] Step S501: Construct a multi-dimensional evaluation vector based on the path length index, average turning angle index, and path complexity index.

[0152] It should be noted that the multidimensional evaluation vector is a vector structure that quantifies path quality using multiple indicators, including path length, average turning angle, and path complexity.

[0153] Taking three-dimensional feature vectors as an example, we can first define the set of ship multi-constraint models as follows: Each multi-constraint model includes the ship's perimeter safety radius. Minimum turning radius of a ship Maximum turning angle of the ship Isokinetic parameters, i.e. Let the set of candidate paths be... Each path corresponds to a three-dimensional feature vector, which is expressed as formula (20).

[0154] (20)

[0155] In formula (20), The quality evaluation index of the generated path is generated by executing the improved RRT-star algorithm. There are 3 paths, and the three-dimensional feature vector of each path corresponds to three sets of evaluation indicators, as expressed by formula (21).

[0156] (twenty one)

[0157] In formula (21), This indicates the path length metric. This represents the average steering angle index. This represents a metric for path complexity. For each path... Construct the original index matrix, which is expressed as formula (22).

[0158] (twenty two)

[0159] In formula (22), The matrix containing the index parameters of each path is used to perform column standardization on the evaluation parameters of the generated path to avoid dimensional differences, as shown in formula (23).

[0160] (twenty three)

[0161] Step S502: Based on the effectiveness penalty mechanism and multi-dimensional evaluation vector, the comprehensive evaluation value of each group of optimized path point sequences is calculated using a weighted comprehensive evaluation algorithm.

[0162] It should be noted that the effectiveness penalty mechanism can be understood as a method for enforcing path compliance. Step S502 includes: calculating intermediate evaluation parameters through the effectiveness penalty mechanism; and calculating the comprehensive evaluation value of each group of optimized path point sequences through a weighted comprehensive evaluation algorithm.

[0163] Specifically, the steps for calculating intermediate evaluation parameters through the effectiveness penalty mechanism include: defining a path identifier vector, which includes effective and invalid states; when an optimized path point sequence violates motion constraints, setting the corresponding identifier vector to an invalid state and defining the optimized path point sequence that violates motion constraints as an invalid path; and multiplying the quality evaluation parameters of the invalid path by a preset penalty coefficient to obtain intermediate evaluation parameters.

[0164] For example, a validity penalty mechanism is set up, and the path validity identifier vector A is defined as formula (24).

[0165] (twenty four)

[0166] In formula (24), This is represented by formula (25).

[0167] (25)

[0168] The penalty imposed on invalid paths is expressed as formula (26).

[0169] (26)

[0170] In formula (26), The purpose of this is to expand the quality assessment of invalid paths by 1.5 times.

[0171] Step S503: Based on the comprehensive evaluation value, the optimal target planning path is obtained.

[0172] It should be noted that the weighted comprehensive evaluation algorithm can assign weights according to the importance of each indicator, and calculate the comprehensive evaluation value of the path by combining the multi-dimensional evaluation vector and the intermediate evaluation parameters adjusted by penalties.

[0173] For example, selecting the optimal path through weighted comprehensive evaluation can be expressed as formula (27).

[0174] (27)

[0175] In formula (27), the weight vector for Then the optimal path index It can be determined by formula (28).

[0176] (28)

[0177] It should be noted that the advantage of the multi-path planning strategy is that the generated path can adapt to different types of ship operating conditions, provide multiple alternatives, increase the success rate of planning, and at the same time, it can be comprehensively evaluated according to actual needs to select the most suitable path.

[0178] In this embodiment, a multi-dimensional evaluation vector is constructed to address the problem that traditional single-index evaluation cannot comprehensively reflect path quality. An effectiveness penalty mechanism is used to apply penalty coefficients to paths that violate motion constraints, generating intermediate evaluation parameters and resolving the defect of invalid paths being included in the evaluation. Combined with a weighted comprehensive evaluation algorithm, the multi-dimensional vector and penalty results are integrated, overcoming the limitation of unreasonable multi-objective weight allocation. Finally, the optimal objective planning path is selected based on the comprehensive evaluation value, solving the problem of randomness in manual experience-based selection, and a global comparison is derived to directly output the globally optimal solution under multiple constraints.

[0179] In one implementation, to verify the effectiveness of the improved RRT-star algorithm in the above embodiments, an experimental environment based on the topological features of the Zhoushan Islands was constructed. The core of this environment lies in achieving coupled analysis of obstacle modeling and ship motion constraints. When constructing the simulation environment, vector polygon modeling technology was used to achieve high-precision marine topography reconstruction. By defining a structure array containing the coordinates of polygon vertices, the irregular contours of the main islands were accurately described.

[0180] Specifically, in terms of motion constraint modeling, this study constructed a ship motion model with five different constraint levels. Table 1 shows the parameters of the ship multi-constraint model. Based on the combined effect of ship navigation safety distance constraints, minimum turning radius constraints, and maximum turning angles, the constraints are divided into five levels: ultra-strict (SS), strict (S), medium (M), lenient (L), and ultra-flexible (UF). Through this hierarchical setting, the system evaluates the algorithm's adaptability to unmanned vessel path planning under different constraint levels.

[0181] Table 1

[0182]

[0183] Table 2 shows the parameter details of the algorithm for USV path planning under different constraints.

[0184] Table 2

[0185]

[0186] Figure 9 The improved RRT algorithm, based on the five ship multi-constraint models in Table 1, shows five planned paths. It can be seen that the directional changes of each feasible path segment smoothly transition to the next segment within the allowable range of its multi-constraint parameters, making them all suitable as tracking paths for unmanned surface vessels (USVs). Finally, the path length, average turning angle, and number of path points are used as fitness indicators for evaluation. The optimal path is the fourth path (strictly constrained by S). This path is smoother, with a total length of 23338.9m, only three heading change points, and an average directional change of approximately 13.4°. It can be selected as the optimal target path for USVs. The visualization results of the ship multi-constraint model indicator parameters can be found in [reference needed]. Figure 10 and Figure 11 .

[0187] To verify the superior performance of the improved RRT-Star algorithm designed in this paper in ship path planning, we compared the algorithm with the traditional RRT algorithm and the conventional RRT-Star algorithm. First, still in the Zhoushan Islands sea area map environment, we set the starting point as (2000m, 2000m) and the ending point as (18000m, 18000m). The basic settings of the algorithm are shown in Table 3.

[0188] Table 3

[0189]

[0190] Through simulation calculations, we can obtain the paths and parameters planned by three algorithms: traditional RRT, conventional RRT-Star, and improved RRT-Star. Figure 12 As shown in Table 4.

[0191] Table 4

[0192]

[0193] The data shows that the improved RRT-Star algorithm shortens the planned flight distance by 8177.68m compared to the traditional RRT algorithm, a reduction of 25.95%, and by 3809.23m compared to the traditional RRT algorithm, a reduction of 14.03%. In terms of path smoothness, the improved RRT-Star algorithm improves performance by 64.69% compared to the traditional RRT algorithm and by 48.22% compared to the traditional RRT algorithm. Path complexity is reduced by 91.46% compared to the traditional RRT algorithm and by 88.1% compared to the traditional RRT algorithm.

[0194] It should be noted that, in order to address the problems of insufficient collaborative design between planning and control in traditional methods, which leads to frequent path switching and large fluctuations in rudder angle commands, as well as the limited robustness of traditional schemes to multi-source disturbances (such as ocean currents and waves) and the lack of adaptability, the following implementation method is proposed.

[0195] In one implementation, based on the above embodiments, after step S50, the method further includes: calculating the desired heading angle based on the optimal target planning path using an adaptive guidance strategy; obtaining the actual heading angle based on the real-time motion state of the ship; generating a rudder angle command using a fuzzy adaptive PID controller based on the deviation between the desired heading angle and the actual heading angle; updating the ship's motion state according to the rudder angle command; and cyclically executing the guidance strategy for path tracking.

[0196] For example, such as Figure 13 As shown, after determining the optimal target planning path, the ship's real-time coordinates can be obtained through the ship motion mathematical model. Then, the ship's actual desired heading angle can be obtained through line-of-sight guidance algorithms or strategies. Then, the ship's real-time heading angle is obtained according to the ship's equation of motion. The heading deviation is obtained by subtracting the observed heading angle from the actual expected heading angle. The heading deviation rate was also calculated. Then adjust this heading deviation Heading deviation rate The parameters are input into the fuzzy adaptive PID control algorithm program. The heading deviation and heading deviation rate are calculated by the fuzzy algorithm and the three gain parameters of the PID are output in real time. Then the output of the fuzzy algorithm is... Gain parameters and heading deviation The rudder angle of the ship can be output after being calculated by the PID controller. The ship's rudder angle is input into the steering gear (which can be calculated using the ship's equations of motion), allowing for the calculation and updating of the ship's motion state at the next moment. This forms a closed loop, continuously updating the ship's state to gradually bring it closer to the target waypoint. Finally, after updating the ship's motion state information, it checks whether the target waypoint has been reached. If not, it returns to the adaptive guidance strategy stage to continue the tracking process; if so, it checks whether the current waypoint is the last point of the planned path. If not, it switches to the next waypoint to continue tracking; if so, the process ends, completing the path tracking task.

[0197] In this embodiment, the problem of control input response lag is solved by acquiring the deviation between the actual and desired heading angles in real time, and the basis for instantaneous correction is derived through accurate deviation feedback. Based on the deviation, a fuzzy adaptive PID controller generates rudder angle commands, addressing the issue of insufficient robustness under multi-source disturbances. Disturbance resistance capability is derived through fuzzy rule-based adaptive parameter tuning (dynamically optimizing PID coefficients based on the deviation and its rate of change). Finally, the ship's state is updated according to the rudder angle commands, and the guidance strategy is executed cyclically, collaboratively achieving a breakthrough in path tracking stability.

[0198] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the path planning method of the unmanned vessel of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0199] This application also provides a path planning device for an unmanned surface vessel, please refer to... Figure 14 The path planning device for the unmanned vessel includes:

[0200] The environmental map creation module 10 is used to create an environmental map based on the ship's current location information, destination location information, ship status information, and obstacle location information.

[0201] The initial path generation module 20 is used to generate an initial path point sequence in the environmental map based on ship motion characteristic constraints and through an improved asymptotic optimal fast exploration random tree algorithm.

[0202] The optimized path generation module 30 is used to perform motion constraint-aware path optimization processing on the initial path point sequence and output an optimized path point sequence. The path optimization processing is used to optimize issues such as path redundancy, frequent turns, and direct flight feasibility.

[0203] The multi-path generation module 40 is used to obtain multiple sets of optimized path point sequences based on ship characteristic parameters with different levels of constraints.

[0204] The target path planning module 50 is used to perform a multi-objective comprehensive evaluation of the path quality of multiple sets of optimized path point sequences, and to select the optimal target planning path.

[0205] The unmanned vessel path planning device provided in this application, employing the unmanned vessel path planning method described in the above embodiments, can solve the technical problems of insufficient feasibility and low efficiency in vessel path planning under complex sea conditions. Compared with the prior art, the beneficial effects of the unmanned vessel path planning device provided in this application are the same as those of the unmanned vessel path planning method provided in the above embodiments, and other technical features in the unmanned vessel path planning device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0206] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the path planning method for the unmanned vessel in the above embodiments.

[0207] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0208] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by the path planning device of the unmanned vessel, the path planning device of the unmanned vessel: establishes an environmental map based on the vessel's current position information, destination position information, vessel status information, and obstacle position information; generates an initial path point sequence in the environmental map based on vessel motion characteristic constraints using an improved asymptotic optimal fast exploration random tree algorithm; performs motion constraint-aware path optimization processing on the initial path point sequence, outputting an optimized path point sequence, wherein the path optimization processing is used to optimize issues such as path redundancy, frequent turning, and direct navigation feasibility; obtains multiple sets of the optimized path point sequences based on vessel characteristic parameters with different levels of constraints; and performs a multi-objective comprehensive evaluation of path quality on the multiple sets of optimized path point sequences, selecting the optimal target planning path.

[0209] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0210] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the aforementioned unmanned vessel path planning method. This addresses the technical problems of insufficient feasibility and low efficiency in vessel path planning under complex sea conditions. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the unmanned vessel path planning method provided in the above embodiments, and will not be elaborated upon here.

[0211] This application provides a path planning device for an unmanned vessel, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the path planning method for the unmanned vessel in the first embodiment described above.

[0212] The following is for reference. Figure 15The diagram illustrates a structural schematic of a path planning device suitable for implementing the embodiments of this application for unmanned vessels. The path planning device for unmanned vessels in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 15 The path planning device for the unmanned vessel shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0213] like Figure 15 As shown, the path planning device of the unmanned surface vessel (USV) may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the USV's path planning device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. The communication device 1009 allows the unmanned vessel's path planning equipment to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 15 The diagram illustrates a path planning device for an unmanned surface vessel with various systems; however, it should be understood that implementation of or possession of all of the systems shown is not required. More or fewer systems may be implemented alternatively.

[0214] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0215] The unmanned vessel path planning device provided in this application, employing the unmanned vessel path planning method described in the above embodiments, can solve the technical problems of insufficient feasibility and low efficiency in vessel path planning under complex sea conditions. Compared with the prior art, the beneficial effects of the unmanned vessel path planning device provided in this application are the same as those of the unmanned vessel path planning method provided in the above embodiments, and other technical features of the unmanned vessel path planning device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0216] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0217] The above description is only part of the embodiments of the present application and does not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A path planning method for an unmanned surface vessel, characterized in that, The method includes: An environmental map is created based on the ship's current location, destination location, ship status, and obstacle location information. In the environmental map, a dual-constraint model of ship motion is constructed, which is used to constrain the maximum steering angle and the minimum turning radius; Random sampling points are generated using a dynamic hybrid sampling mechanism, which includes path enhancement sampling, target bias sampling, heuristic sampling, and uniform random sampling. The random sampling points are expanded to obtain an expanded node set; The ship motion dual-constraint model is applied to the extended node set, and collision detection and steering feasibility verification are performed on each node to iteratively generate an initial path point sequence. The initial path point sequence is subjected to motion constraint-aware path optimization processing to output an optimized path point sequence. The path optimization processing is used to optimize issues such as path redundancy, frequent turns, and direct flight feasibility. Multiple sets of optimized path point sequences are obtained based on ship characteristic parameters with different levels of constraints; A multi-objective comprehensive evaluation of path quality is performed on multiple sets of optimized path point sequences to select the optimal objective planning path. The step of performing motion constraint-aware path optimization processing on the initial path point sequence and outputting an optimized path point sequence includes: The initial optimized path points are obtained by removing intermediate nodes that meet the redundancy condition from the initial path point sequence through path pruning. By performing a skip connection on the initial optimized path points, intermediate optimized path points are obtained; Based on the redundancy point geometric elimination criterion, the intermediate optimized path points are optimized to obtain an optimized path point sequence. The redundancy point geometric elimination criterion, for path points after skip connections, deletes or adjusts points that do not conform to the straight-line continuity of the ship based on the geometric constraint of the minimum turning radius, thereby generating the optimized path point sequence. The step of performing a multi-objective comprehensive evaluation of path quality on multiple sets of optimized path point sequences and selecting the optimal objective planning path includes: A multi-dimensional evaluation vector is constructed based on path length, average turning angle, and path complexity metrics. Based on the effectiveness penalty mechanism and the multidimensional evaluation vector, the comprehensive evaluation value of each group of optimized path point sequences is calculated by a weighted comprehensive evaluation algorithm; The optimal target planning path is obtained by filtering based on the comprehensive evaluation value.

2. The method as described in claim 1, characterized in that, The step of calculating the comprehensive evaluation value of each group of optimized path point sequences using a weighted comprehensive evaluation algorithm based on the effectiveness penalty mechanism and the multi-dimensional evaluation vector includes: Intermediate evaluation parameters are calculated using an effectiveness penalty mechanism; The comprehensive evaluation value of each group of optimized path point sequences is calculated using a weighted comprehensive evaluation algorithm.

3. The method as described in claim 2, characterized in that, The step of calculating intermediate evaluation parameters through the effectiveness penalty mechanism includes: Define a path identifier vector, which includes valid and invalid states; When the optimized path point sequence violates the motion constraints, the corresponding identifier vector is set to an invalid state, and the optimized path point sequence that violates the motion constraints is defined as an invalid path; The intermediate evaluation parameters are obtained by multiplying the quality evaluation parameters of invalid paths by the preset penalty coefficients.

4. The method according to any one of claims 1 to 3, characterized in that, After the step of performing a multi-objective comprehensive evaluation of path quality on multiple sets of optimized path point sequences and selecting the optimal objective planning path, the method further includes: Based on the optimal target planning path, the desired heading angle is calculated using an adaptive guidance strategy; The actual heading angle is obtained based on the ship's real-time motion status; Based on the deviation between the desired heading angle and the actual heading angle, a fuzzy adaptive PID controller generates a rudder angle command. The ship's motion state is updated according to the rudder angle command, and the guidance strategy is executed cyclically for path tracking.

5. A path planning device for an unmanned surface vessel, characterized in that, The device includes: The environmental map creation module is used to create an environmental map based on the ship's current location information, destination location information, ship status information, and obstacle location information. An initial path generation module is used to construct a dual-constraint model of ship motion in the environmental map, which constrains the maximum steering angle and the minimum turning radius; it generates random sampling points using a dynamic hybrid sampling mechanism, which includes path enhancement sampling, target bias sampling, heuristic sampling, and uniform random sampling; it expands the random sampling points to obtain an expanded node set; it applies the dual-constraint model of ship motion to the expanded node set, performs collision detection and steering feasibility verification on each node, and iteratively generates an initial path point sequence; The optimized path generation module is used to perform motion constraint-aware path optimization processing on the initial path point sequence and output an optimized path point sequence. The path optimization processing is used to optimize issues such as path redundancy, frequent turns, and direct flight feasibility. Multiple path generation modules are used to obtain multiple sets of optimized path point sequences based on ship characteristic parameters with different levels of constraints; The target path planning module is used to perform a multi-objective comprehensive evaluation of the path quality of multiple sets of optimized path point sequences, and to select the optimal target planning path. The optimized path generation module is further configured to remove intermediate nodes that meet the redundancy condition in the initial path point sequence through path pruning to obtain initial optimized path points; to perform skip connections on the initial optimized path points to obtain intermediate optimized path points; and to optimize the intermediate optimized path points based on the redundancy point geometric elimination criterion to obtain an optimized path point sequence. The redundancy point geometric elimination criterion, for the path points after skip connections, deletes or adjusts points that do not conform to the continuity of straight-line navigation of the ship based on the geometric constraint of the minimum turning radius, thereby generating the optimized path point sequence. The target path planning module is further configured to construct a multi-dimensional evaluation vector based on path length, average turning angle, and path complexity indicators; calculate the comprehensive evaluation value of each group of optimized path point sequences using a weighted comprehensive evaluation algorithm based on the effectiveness penalty mechanism and the multi-dimensional evaluation vector; and obtain the optimal target planning path by filtering based on the comprehensive evaluation value.

6. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the path planning method for the unmanned vessel as described in any one of claims 1 to 4.

7. A path planning device for an unmanned surface vessel, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the path planning method for an unmanned vessel as described in any one of claims 1 to 4.

Images