Marine unmanned aerial vehicle path planning method

Abstract

The invention discloses a maritime unmanned aerial vehicle path planning method, which belongs to the technical field of path planning and comprises the steps of defining a three-dimensional grid map, preliminarily planning a path, optimizing the path and navigating the path. According to the scheme, bidirectional search is carried out through double search trees, the dynamic offset probability is calculated according to the obstacle density, the sampling point is generated, the nearest neighbor node is found, the gravitation of the target point, the repulsive force of the obstacle and the environmental force of the sea wind are combined to obtain the expansion resultant force, and then the adaptive expansion step length is obtained by combining the obstacle density. New nodes are generated and collision detection is carried out, and double search trees are connected to obtain an initial path, so that the efficiency, adaptability and safety of path planning are remarkably improved; according to the method, the Euclidean distance, the energy consumption value and the environment threat value are weighted and combined to obtain a comprehensive evaluation value of a path segment, an optimal jump node is selected for pruning, cubic spline interpolation smoothing processing is carried out to obtain a final path, and the flight requirement of the marine unmanned aerial vehicle can be better met.

Description

Technical Field

[0001] The present invention belongs to the technical field of path planning, and specifically relates to a method for planning a path for a maritime unmanned aerial vehicle (UAV). Background Art

[0002] Maritime UAV path planning methods are based on artificial intelligence technology and provide safe and efficient flight trajectories for autonomous navigation of UAVs in complex marine environments. This ensures that UAVs can complete their missions safely and efficiently in complex and changing marine environments, and provides accurate path planning support for fields such as marine monitoring, marine rescue, and marine resource exploration, thereby improving the efficiency and safety of marine operations. However, existing maritime UAV path planning methods suffer from the complex and changing marine environment, uneven distribution of obstacles, and dynamic changes. Traditional path planning methods have difficulty efficiently handling dynamic constraints in complex environments, resulting in low path planning efficiency, poor path adaptability, and insufficient safety. Existing maritime UAV path planning methods also have the problem that the planned paths may contain redundant nodes, resulting in high flight energy consumption, frequent maneuvers, and insufficient path smoothness, affecting the stability and flight efficiency of the UAV. Summary of the Invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a method for path planning of marine UAVs. In view of the complex and changeable marine environment, uneven and dynamic obstacle distribution in the existing marine UAV path planning methods, traditional path planning methods are difficult to efficiently handle dynamic constraints in complex environments, resulting in low path planning efficiency, poor path adaptability and insufficient safety. This solution performs two-way search through dual search trees to speed up path planning; calculates dynamic bias probability according to obstacle density, generates sampling points and finds nearest neighbor nodes, thereby improving the flexibility of path planning; combines the gravitational force of the target point, the repulsive force of the obstacle and the environmental force of the sea breeze to obtain an extended resultant force, simulates the actual force conditions of the marine UAV, and makes path planning more in line with actual flight requirements; and obtains an adaptive extension step length in combination with the obstacle density to improve the quality of the path and balance efficiency and safety; generates new nodes and performs collision detection to effectively reduce the risk of collision; The dual search tree connection obtains the preliminary path, which significantly improves the efficiency, adaptability and safety of path planning. In view of the problem that the planned path of existing marine UAV path planning methods may contain redundant nodes, resulting in high flight energy consumption, frequent maneuvers, and insufficient path smoothness, which affects the stability and flight efficiency of the UAV, this scheme extracts an ordered node sequence of the preliminary path to provide a structured data basis for subsequent evaluation. The Euclidean distance, energy consumption value and environmental threat value are weighted and combined to obtain a comprehensive evaluation value of the path segment, which can comprehensively evaluate the pros and cons of the path segment. The Euclidean distance can optimize the length of the path, the consideration of the energy consumption value can optimize the energy efficiency of the path, and the introduction of the environmental threat value can effectively evaluate the safety of the path segment. The optimal jump node is selected for pruning, and then cubic spline interpolation smoothing is performed to obtain the final path, which reduces maneuvering energy consumption, improves the flight stability, controllability and flight efficiency of the UAV, and can better meet the flight needs of marine UAVs.

[0004] The technical solution adopted by the present invention is as follows: The present invention provides a method for planning a path for a marine drone, the method comprising the following steps:

[0005] Step S1: defining a three-dimensional grid map;

[0006] Step S2: preliminary path planning;

[0007] Step S3: path optimization;

[0008] Step S4: route navigation.

[0009] Furthermore, in step S1, the three-dimensional grid map is defined by establishing a three-dimensional coordinate system O-xyz with the starting point of the UAV as the origin, and collecting the marine UAV data based on the three-dimensional coordinate system;

[0010] The data of the maritime drone includes marine environment data and the target points of the drone;

[0011] The marine environment data includes sea area spatial range data, obstacle data, and sea breeze data;

[0012] Based on the sea area spatial range data, a three-dimensional grid map is constructed. Based on the obstacle data, static obstacles and dynamic obstacles are marked in the three-dimensional grid map, and the starting point and target point of the drone are marked.

[0013] Furthermore, in step S2, the preliminary path planning is to plan the path of the maritime drone based on a double search tree to obtain a preliminary path; it specifically includes the following steps:

[0014] Step S21: Set the root node; use the starting point and target point of the drone as the root nodes of the forward search tree T1 and the backward search tree T2 respectively;

[0015] Step S22: Dynamic biasing sampling strategy; calculate the dynamic biasing probability q according to the obstacle density k, and generate a random number s within the range of [0, 1]; if s < q, directly use the target point of the drone as the sampling point; otherwise, randomly generate a sampling point in the free space; find a node in T1 and T2 that is closest to the sampling point Q samp as the nearest neighbor node Q near , and record the samp Euclidean distance between Q near and Q ; the formula used is as follows:

[0016] ;

[0017] ;

[0018] In the formula, F obs and F free are the obstacle occupancy volume and the free space volume respectively;

[0019] Step S23: Calculate the extended resultant force; combine the gravitational force of the target point, the repulsive force of the obstacles, and the environmental force of the sea breeze to obtain the extended resultant force; it includes the following steps:

[0020] Step S231: Calculate the gravitational force; use the product of the Euclidean distance and the gravitational gain coefficient as the gravitational force C near of the target point on Q gra ;

[0021] Step S232: Calculate the repulsive force; for each obstacle, if the Euclidean distance between the center point of obstacle I a and Q near is greater than the maximum influence range D0 of the repulsive force, then Ia Q near The repulsion is 0; otherwise, based on the Euclidean distance and repulsion gain coefficient, the product of the basic repulsion strength, target distance weight and direction vector is used as I a Q near Repulsion; move all obstacles to Q near The repulsive forces are superimposed to obtain the total repulsive force C rep ;

[0022] Step S233: Calculate the environmental force; based on the wind speed and wind direction of the sea breeze, combined with the environmental force gain coefficient L env , we can get the effect of ocean environment on Q near Environmental force C env ;

[0023] Step S234: Synthesize the expansion force; synthesize the attraction, repulsion and environmental forces on Q near The influence of the

[0024] Step S24: Generate adaptive expansion step length; set the maximum expansion step length R max ,based on and R max , combined with the obstacle density k, the adaptive expansion step size R is obtained;

[0025] Step S25: Generate a new node; according to Q near , R and C join Calculate the new node Q new ; Detect the generated Q new Is it within the three-dimensional space? If so, then Q new Is a valid node, go to step S26; otherwise, Q new If it is an invalid node, go to step S22;

[0026] Step S26: Collision detection; Based on the linear motion model, predict the UAV from Q near Xiang Q new The positions of all dynamic obstacles during the movement are detected when the drone moves from Q near Xiang Q new Will it collide with obstacles during the movement? If not, Q new As Q near The child nodes are added to Q near The search tree to which it belongs is found, and the process goes to step S27; otherwise, if there is a collision, the generated Q is discarded. new , and go to step S22;

[0027] Step S27: Double search tree connection; new Each node in another search tree is different, predicting the drone from Qnew Xiang Q f The positions of all dynamic obstacles during the movement are detected when the drone moves from Q new Xiang Q f Will it collide with obstacles during the movement? If there is a node Q f , so that the drone does not collide with obstacles, then splice Q new The path to the root node of the search tree to which it belongs and Q f The path to the root node of the search tree to which it belongs, and the candidate path J is obtained f ; Then from all candidate paths J f Select a shortest path as the initial path J first , complete the preliminary path planning; otherwise, go to step S22; where Q f Yes and Q new The fth node in another different search tree, f is the node index, J f Based on Q f Generated candidate paths.

[0028] Furthermore, in step S3, the path optimization is to optimize the preliminary path by pruning and path smoothing to obtain the final path; specifically, the following steps are included:

[0029] Step S31: Initialization; from the initial path J first Extract the ordered node sequence {Q1, Q2, ..., Q U}, and create an empty list E to store the optimized node sequence, initialize the current node index i=1; where U is the preliminary path J first The total number of nodes in Q1, Q2 and Q U The initial path J first The first, second and Uth nodes in Q1 and Q U They correspond to the starting point and target point of the drone respectively;

[0030] Step S32: evaluating and pruning nodes segment by segment; including the following steps:

[0031] Step S321: Evaluate nodes segment by segment; from the current node Q i Start and traverse all subsequent nodes Q j , j increases from i+1 to U, the energy consumption value is obtained by combining the air resistance energy consumption and the maneuvering energy consumption, the environmental threat value is obtained by combining the obstacle threat and the wind speed threat, and the weighted combination of the Euclidean distance, the energy consumption value and the environmental threat value is obtained. i to Q j The comprehensive evaluation value of this path;

[0032] Step S322: Pruning; for each node Qj , predicting drones from Q i Xiang Q j The positions of all dynamic obstacles during the movement are detected when the drone moves from Q i Xiang Q j Will it collide with obstacles during the movement? If there is a node Q j , so that the drone does not collide with obstacles, then Q j As a candidate node, the node with the smallest comprehensive evaluation value is selected from all candidate nodes as the optimal jump node Q best , Q best and Q i Add to E and update i to Q best Otherwise, only Q i Add to E;

[0033] Step S323: Determine the simplified path; if i=U, connect the nodes in list E in order to obtain the simplified path; otherwise, if i=U-1, connect Q U Add it to E, and then connect the nodes in list E in order to obtain a simplified path; otherwise, add 1 to i and go to step S321 for the next iteration;

[0034] Step S33: path smoothing; performing cubic spline interpolation smoothing processing on the simplified path to obtain the final path, thus completing the path optimization.

[0035] Furthermore, in step S4, the path navigation is to load the optimized final path into the navigation control system of the UAV. Under the guidance of the navigation control system, the UAV flies from the starting point to the target point according to the loaded path information and the final path.

[0036] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0037] (1) In view of the fact that the existing marine environment is complex and changeable, and obstacles are unevenly distributed and dynamically changing, traditional path planning methods are difficult to efficiently handle dynamic constraints in complex environments, resulting in low path planning efficiency, poor path adaptability and insufficient safety. This scheme uses a dual search tree to perform bidirectional search, significantly reducing the number of search iterations and accelerating the path planning speed; the dynamic bias probability is calculated based on the obstacle density, sampling points are generated and the nearest neighbor nodes are found, thereby improving the flexibility of path planning; the gravity of the target point, the repulsive force of the obstacle and the environmental force of the sea breeze are combined to obtain the extended force, simulating the actual force situation of the marine UAV, adapting to marine environments of different complexities, and making the path planning more in line with actual flight requirements; the obstacle density is combined to obtain an adaptive extension step size, generating a more refined path, improving the quality of the path, and balancing efficiency and safety; new nodes are generated and collision detection is performed, effectively reducing the collision risk; the dual search trees are connected to obtain a preliminary path, which significantly improves the efficiency, adaptability and safety of path planning.

[0038] (2) In view of the problem that the existing path planning methods for marine UAVs may contain redundant nodes in the planned path, resulting in high flight energy consumption, frequent maneuvers, and insufficient path smoothness, which affects the stability and flight efficiency of the UAV, this scheme extracts an ordered node sequence of the preliminary path to provide a structured data basis for subsequent evaluation and avoid optimization deviation caused by node disorder; the Euclidean distance, energy consumption value and environmental threat value are weighted and combined to obtain a comprehensive evaluation value of the path segment, which can comprehensively evaluate the pros and cons of the path segment. The Euclidean distance can optimize the length of the path, the consideration of the energy consumption value can optimize the energy efficiency of the path, reduce the flight energy consumption of the UAV, and extend the flight time. The introduction of the environmental threat value can effectively evaluate the safety of the path segment, avoid the UAV from entering the high-threat area, and improve flight safety; the optimal jump node is selected for pruning, and then the cubic spline interpolation smoothing process is performed to obtain the final path, which reduces the frequent turning and maneuvering of the UAV, reduces the maneuvering energy consumption, improves the flight stability, controllability and flight efficiency of the UAV, and can better meet the flight needs of marine UAVs. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A schematic diagram of a flow chart of a method for planning a path for a maritime UAV provided by the present invention;

[0040] Figure 2 Schematic diagram of the process of step S2;

[0041] Figure 3 Schematic diagram of the process of step S3.

[0042] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0044] In the description of the present invention, it should be understood that terms such as "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they should not be understood as limiting the present invention.

[0045] Example 1, see Figure 1 The present invention provides a method for planning a path for a maritime UAV, which comprises the following steps:

[0046] Step S1: define a three-dimensional grid map; construct a three-dimensional grid map and mark obstacles, starting points, and target points;

[0047] Step S2: Preliminary path planning: Perform a bidirectional search using a dual search tree, calculate the dynamic bias probability based on the obstacle density, generate sampling points, and find the nearest neighbor nodes. Combine the attraction of the target point, the repulsion of the obstacle, and the environmental force of the sea breeze to obtain the expansion resultant force. Combine this with the obstacle density to obtain an adaptive expansion step size. Generate new nodes and perform collision detection. Connect the dual search trees to obtain a preliminary path.

[0048] Step S3: Path optimization: extract the ordered node sequence of the preliminary path, combine the Euclidean distance, energy consumption value and environmental threat value in a weighted manner to obtain the comprehensive evaluation value of the path segment, select the optimal jump node for pruning, and then perform cubic spline interpolation smoothing to obtain the final path;

[0049] Step S4: Path navigation; the drone flies from the starting point to the target point according to the final path.

[0050] Example 2, see Figure 1This embodiment is based on the above embodiment. In step S1, the three-dimensional grid map is defined by establishing a three-dimensional coordinate system O-xyz with the starting point of the drone as the origin, wherein the x-axis points to the east, the y-axis points to the north, and the z-axis is vertically upward, and the marine drone data is collected based on the three-dimensional coordinate system;

[0051] The marine drone data includes marine environment data and the target point of the drone;

[0052] The marine environment data includes sea area spatial range data, obstacle data and sea wind data;

[0053] The sea area spatial range data includes the horizontal coordinate range and the maximum flight altitude, and the three-dimensional spatial range is obtained. ; where x min and x max are the minimum and maximum values of the x-axis, and the y- min and y max are the minimum and maximum values of the y-axis, and the z max is the maximum value of the z-axis;

[0054] The obstacle data includes static obstacle data and dynamic obstacle data;

[0055] The static obstacle data includes the coordinates of the center point of the static obstacle and the radius of the static obstacle;

[0056] The dynamic obstacle data includes the coordinates of the dynamic obstacle center point, the dynamic obstacle radius, the dynamic obstacle heading angle and the dynamic obstacle speed;

[0057] The sea wind data includes wind speed and wind direction angle;

[0058] A three-dimensional grid map is constructed based on the spatial range data of the sea area. Static obstacles and dynamic obstacles are marked in the three-dimensional grid map based on the obstacle data, and the starting point and target point of the UAV are marked.

[0059] Example 3, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S2, the preliminary path planning is to perform path planning for the maritime UAV based on a dual search tree to obtain a preliminary path. Specifically, the following steps are included:

[0060] Step S21: Set the root node; Traditional single-tree search expands unidirectionally from the starting point to the ending point, which is inefficient in complex marine environments. By using bidirectional search to accelerate path discovery, it can meet in the middle area faster, reduce the number of search iterations, is suitable for long-distance path planning scenarios at sea, saves computing time, and reduces the search complexity in large-scale three-dimensional spaces; Take the starting point and the target point of the UAV as the root nodes of the forward search tree T1 and the backward search tree T2 respectively;

[0061] Step S22: Dynamic biasing sampling strategy; The distribution of obstacles at sea is uneven, and the fixed sampling strategy cannot balance goal orientation and free space exploration. Dynamically adjust the sampling probability according to the obstacle density. Increase the sampling probability of the target point when the obstacles are dense to guide the search to quickly point to the target. Randomly sample when the proportion of free space is large to avoid getting stuck in local optima, which can adapt to different complex marine environments and improve the flexibility of path planning; Calculate the dynamic biasing probability q according to the obstacle density k, and generate a random number s within the range of [0, 1]; If s < q, directly take the target point of the UAV as the sampling point; Otherwise, randomly generate a sampling point within the free space; Find a node in T1 and T2 that is closest to the sampling point Q samp and take it as the nearest neighbor node Q near and record the samp Euclidean distance between Q near and Q ; The formula used is as follows:

[0062] ;

[0063] ;

[0064] In the formula, F obs and F free are the occupied volume of obstacles and the free space volume respectively. The occupied volume of obstacles includes the occupied volume of static obstacles and the occupied volume of dynamic obstacles. The free space volume is the total sea area volume in the three-dimensional grid map minus the occupied volume of obstacles;

[0065] Step S23: Calculate the expansion resultant force; The flight of a UAV at sea is affected by target attraction, obstacle repulsion, and sea breeze. It is necessary to construct a multi-force coupling model to guide the path expansion direction, quantify the dynamic impact of environmental factors on the flight trajectory, and consider target orientation, obstacle avoidance, and sea breeze interference at the same time, which meets the flight requirements in complex marine environments; Combine the gravitational force of the target point, the repulsive force of the obstacles, and the environmental force of the sea breeze to obtain the expansion resultant force; It includes the following steps:

[0066] Step S231: Calculate the gravitational force; Guide the UAV to approach the target point, avoid the search direction from deviating, and reduce ineffective search; Take the product of the Euclidean distance and the gravitational gain coefficient as the gravitational force of the target point on Q nearGravity C gra ; The formula used is as follows:

[0067] ;

[0068] Where, L gra is the gravitational gain coefficient, Q tar is the target point of the drone, It's Q tar and Q near The Euclidean distance between

[0069] Step S232: Calculate repulsion; guide the drone away from obstacles, avoid collision risks in advance, balance obstacle avoidance and target guidance, and avoid excessive obstacle avoidance resulting in a long path; for each obstacle, if the obstacle I a The center point and Q near The Euclidean distance between them is greater than the maximum influence range D0 of the repulsive force, then I a Q near The repulsion is 0; otherwise, based on the Euclidean distance and repulsion gain coefficient, the product of the basic repulsion strength, target distance weight and direction vector is used as I a Q near Repulsion; move all obstacles to Q near The repulsive forces are superimposed to obtain the total repulsive force C rep ; The formula used is as follows:

[0070] ;

[0071] Where C rep is all obstacles to Q near The total repulsive force, I a is the ath obstacle, isI a Q near The repulsive force, L rep is the repulsion gain coefficient, isI a The center point, yes and Q near The Euclidean distance between 、 and They are the basic repulsive strength, target distance weight and direction vector;

[0072] Step S233: Calculate environmental forces; sea breeze is a unique environmental interference factor at sea, which will affect the flight speed and energy consumption of the UAV, and its impact on the path needs to be quantified; according to the wind speed and direction of the sea breeze, combined with the environmental force gain coefficient L env , we can get the effect of ocean environment on Q near Environmental force Cenv ; The formula used is as follows:

[0073] ;

[0074] Where C env Is the marine environment to Q near Environmental force, v b and θ b are the wind speed and wind direction angle of the sea breeze respectively;

[0075] Step S234: Synthesize and extend the combined force; the multi-force synthesis model simulates the actual force situation of the UAV at sea, so that the path planning can meet the goal orientation, obstacle avoidance requirements and environmental adaptability at the same time, and improve the feasibility and safety of the path; the comprehensive gravity, repulsion and environmental force on Q near The expansion resultant force is obtained by the following formula:

[0076] ;

[0077] Where C join The gravitational force, repulsive force and environmental force on Q near The expansion force;

[0078] Step S24: Generate an adaptive expansion step size; when the density of obstacles at sea changes, a fixed step size cannot take into account both efficiency and safety. In areas with dense obstacles, the step size is automatically reduced to increase the precision of path planning and reduce the risk of collision. In open sea areas, the step size is increased to speed up the search and balance efficiency and safety. The step size is dynamically associated with the obstacle density to adapt to the randomness of the marine environment. Set the maximum expansion step size R max ,based on and R max , combined with the obstacle density k, the adaptive expansion step length R is obtained; the formula used is as follows:

[0079] ;

[0080] Step S25: Generate a new node; node generation combines the direction of the resultant force and the adaptive step size to ensure that the path is extended along the optimal direction and that the new node is located within the sea area to avoid planning a path that exceeds the flight boundary of the drone; according to Q near , R and C join Calculate the new node Q new ; Detect the generated Q new Is it within the three-dimensional space? If so, then Q new Is a valid node, go to step S26; otherwise, Q new If it is an invalid node, go to step S22; the formula used is as follows:

[0081] ;

[0082] Step S26: Collision detection; by predicting the trajectory of obstacles, avoid collision risks in advance and ensure path feasibility; based on the linear motion model, predict the UAV from Q near Xiang Q new The positions of all dynamic obstacles during the movement are detected when the drone moves from Q near Xiang Q new Will it collide with obstacles during the movement? If not, Q new As Q near The child nodes are added to Q near The search tree to which it belongs is found, and the process goes to step S27; otherwise, if there is a collision, the generated Q is discarded. new , and go to step S22;

[0083] Step S27: Double search tree connection; through bidirectional search and shortest path screening, a more compact path is generated in a complex marine environment, the flight distance and energy consumption are reduced, and a feasible and shortest preliminary path is efficiently generated; new Each node in another search tree is different, predicting the drone from Q new Xiang Q f The positions of all dynamic obstacles during the movement are detected when the drone moves from Q new Xiang Q f Will it collide with obstacles during the movement? If there is a node Q f , so that the drone does not collide with obstacles, then splice Q new The path to the root node of the search tree to which it belongs and Q f The path to the root node of the search tree to which it belongs, and the candidate path J is obtained f ; Then from all candidate paths J f Select a shortest path as the initial path J first , complete the preliminary path planning; otherwise, go to step S22; where Q f Yes and Q new The fth node in another different search tree, f is the node index, J f Based on Q f Generated candidate paths.

[0084] By performing the above operations, existing maritime UAV path planning methods face the challenges of complex and changeable marine environments, unevenly distributed obstacles, and dynamic changes. Traditional path planning methods are unable to efficiently handle dynamic constraints in complex environments, resulting in low path planning efficiency, poor path adaptability, and insufficient safety. This solution uses a dual search tree to perform a bidirectional search, significantly reducing the number of search iterations and accelerating path planning. Dynamic bias probabilities are calculated based on obstacle density to generate sampling points and find nearest neighbor nodes, improving path planning flexibility. The gravitational force of the target point, the repulsive force of obstacles, and the environmental force of the sea breeze are combined to generate an extended resultant force, simulating the actual force conditions of maritime UAVs. This approach adapts to marine environments of varying complexity and makes path planning more consistent with actual flight requirements. The obstacle density is then combined to obtain an adaptive extension step size, generating a more refined path, improving path quality, and balancing efficiency and safety. New nodes are generated and collision detection is performed, effectively reducing collision risk. A preliminary path is obtained by connecting the dual search trees, significantly improving the efficiency, adaptability, and safety of path planning.

[0085] Example 4, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S3, the path optimization is to optimize the preliminary path by pruning and path smoothing to obtain the final path. Specifically, the following steps are included:

[0086] Step S31: Initialization; from the initial path J first Extract the ordered node sequence {Q1, Q2, ..., Q U}, and create an empty list E to store the optimized node sequence, initialize the current node index i=1; where U is the preliminary path J first The total number of nodes in Q1, Q2 and Q U The initial path J first The first, second and Uth nodes in Q1 and Q U They correspond to the starting point and target point of the drone respectively;

[0087] Step S32: Node evaluation and pruning for each segment. The preliminary path may contain redundant nodes, resulting in high flight energy consumption and frequent maneuvers. The pros and cons of each segment are quantified by combining Euclidean distance, energy consumption value, and environmental threat value. The pruning operation selects the optimal jump node through comprehensive evaluation while ensuring collision-free operation, deletes redundant nodes, shortens the path length, reduces flight time, generates a streamlined path, and optimizes the path node sequence to reduce energy consumption and threats. The process includes the following steps:

[0088] Step S321: Evaluate node by node; from the current node Q i Start and traverse all subsequent nodes Q j, j increases from i+1 to U, the energy consumption value is obtained by combining the air resistance energy consumption and the maneuvering energy consumption, the environmental threat value is obtained by combining the obstacle threat and the wind speed threat, and the weighted combination of the Euclidean distance, the energy consumption value and the environmental threat value is obtained. i to Q j The comprehensive evaluation value of this path; the formula used is as follows:

[0089] ;

[0090] ;

[0091] ;

[0092] Where Q i and Q j The initial path J first The i-th and j-th nodes in 、 、 and The drones are from Q i to Q j The Euclidean distance, energy consumption, environmental threat value and comprehensive assessment value of this path, ρ is the air density, V d and W f They are the drag coefficient of the UAV and the frontal area of the UAV, v o is the drone speed, is the wind speed of the sea breeze at path position c, is the curvature of the path at position c, dc is the path element length, γ is the maneuver energy consumption coefficient, A is the total number of obstacles, p obs and p env They are obstacle threat coefficient and environmental threat coefficient, is the safe wind speed threshold, is the path from position c to obstacle I a The Euclidean distance of the center point, r a It is an obstacle I a The radius of the threat field, σ is the attenuation coefficient of the threat field, D max 、B max and G max are the maximum Euclidean distance, maximum energy consumption value, and maximum environmental threat value of all path segments to be evaluated, respectively. ω1, ω2, and ω3 are the weights of Euclidean distance, energy consumption value, and environmental threat value, respectively. ω1=0.5, ω2=0.3, ω3=0.2;

[0093] Step S322: Pruning; for each node Q j , predicting drones from Q i Xiang Q jThe positions of all dynamic obstacles during the movement are detected when the drone moves from Q i Xiang Q j Will it collide with obstacles during the movement? If there is a node Q j , so that the drone does not collide with obstacles, then Q j As a candidate node, the node with the smallest comprehensive evaluation value is selected from all candidate nodes as the optimal jump node Q best , Q best and Q i Add to E and update i to Q best Otherwise, only Q i Add to E;

[0094] Step S323: Determine the simplified path; the structured termination condition ensures that the pruning process is completed efficiently and avoids infinite iterations. The generated simplified path maximizes the optimization path efficiency while ensuring safety, meeting the real-time requirements of marine drones; if i=U, then connect the nodes in list E in order to obtain the simplified path; otherwise, if i=U-1, then Q U Add it to E, and then connect the nodes in list E in order to obtain a simplified path; otherwise, add 1 to i and go to step S321 for the next iteration;

[0095] Step S33: Path smoothing; The streamlined path may have sharp inflection points, which will increase the maneuvering energy consumption and reduce the stability of the UAV during flight. Smoothing can improve the trajectory continuity and path smoothness. Smoothing the trajectory reduces the maneuverability of the UAV and reduces mechanical losses. It can also improve flight stability in strong winds at sea. The streamlined path is smoothed using cubic spline interpolation to obtain the final path, completing the path optimization.

[0096] By performing the above operations, the existing path planning methods for maritime UAVs may contain redundant nodes in the planned path, resulting in high flight energy consumption, frequent maneuvers, and insufficient path smoothness, which affects the stability and flight efficiency of the UAV. This solution extracts an ordered node sequence of the preliminary path to provide a structured data basis for subsequent evaluation and avoid optimization deviations caused by node disorder. The Euclidean distance, energy consumption value, and environmental threat value are weighted and combined to obtain a comprehensive evaluation value of the path segment, which can comprehensively evaluate the quality of the path segment. The Euclidean distance can optimize the length of the path, the consideration of the energy consumption value can optimize the energy efficiency of the path, reduce the flight energy consumption of the UAV, and extend the flight time. The introduction of the environmental threat value can effectively evaluate the safety of the path segment, prevent the UAV from entering high-threat areas, and improve flight safety. The optimal jump node is selected for pruning, and then cubic spline interpolation smoothing is performed to obtain the final path. This reduces the UAV's frequent turning and maneuvering, reduces maneuvering energy consumption, improves the UAV's flight stability, controllability, and flight efficiency, and can better meet the flight needs of maritime UAVs.

[0097] Example 5, see Figure 1 This embodiment is based on the above embodiment. In step S4, the path navigation is to load the optimized final path into the navigation control system of the UAV. Under the guidance of the navigation control system, the UAV flies from the starting point to the target point according to the loaded path information and the final path.

[0098] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0099] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

[0100] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, designs structures and embodiments similar to this technical solution without inventiveness, they shall fall within the scope of protection of the present invention.

Claims

1. A method for planning a path for a maritime UAV, characterized by: The method includes the following steps: Step S1: Define a three-dimensional grid map; construct a three-dimensional grid map, and mark obstacles, starting points, and target points; Step S2: Preliminary path planning; perform bidirectional search through a double search tree, calculate the dynamic bias probability according to the obstacle density, generate sampling points and find the nearest neighbor nodes, combine the gravitational force of the target point, the repulsive force of the obstacles, and the environmental force of the sea breeze to obtain the extended combined force, and then combine with the obstacle density to obtain the adaptive extended step size, generate new nodes and perform collision detection, and connect the double search tree to obtain a preliminary path; Step S3: Path optimization; extract the ordered node sequence of the preliminary path, combine the Euclidean distance, energy consumption value, and environmental threat value with weights to obtain the comprehensive evaluation value of the path segment, select the optimal jump node for pruning, and then perform cubic spline interpolation smoothing to obtain the final path; Step S4: Path navigation; the unmanned aerial vehicle flies from the starting point to the target point according to the final path.

2. A method for planning a path for a maritime UAV according to claim 1, characterized in that: In step S2, the preliminary path planning is to perform path planning for the maritime unmanned aerial vehicle based on a double search tree to obtain a preliminary path; specifically, it includes the following steps: Step S21: Set the root nodes; take the starting point and target point of the unmanned aerial vehicle as the root nodes of the forward search tree T1 and the reverse search tree T2 respectively; Step S22: Dynamic bias sampling strategy; Step S23: Calculate the extended combined force; combine the gravitational force of the target point, the repulsive force of the obstacles, and the environmental force of the sea breeze to obtain the extended combined force; Step S24: Generate adaptive expansion step length; set the maximum expansion step length R max ,based on and R max , combined with the obstacle density k, the adaptive expansion step size R is obtained; where, is the sampling point Q samp and its nearest neighbor node Q near The Euclidean distance between Step S25: Generate a new node; according to Q near , R and C join Calculate the new node Q new ; Detect the generated Q new Is it within the three-dimensional space? If so, then Q new Is a valid node, go to step S26; otherwise, Q new Is an invalid node, go to step S22; wherein, C join The gravitational force, repulsive force and environmental force on Q near The expansion force; Step S26: Collision detection; Based on the linear motion model, predict the UAV from Q near Xiang Q new The positions of all dynamic obstacles during the movement are detected when the drone moves from Q near Xiang Q new Will it collide with obstacles during the movement? If not, Q new As Q near The child nodes are added to Q near The search tree to which it belongs is found, and the process goes to step S27; otherwise, if there is a collision, the generated Q is discarded. new , and go to step S22; Step S27: Connect the double search tree.

3. A method for planning a path for a maritime UAV according to claim 2, characterized in that: In step S22, the dynamic bias sampling strategy is to calculate the dynamic bias probability q according to the obstacle density k, and generate a random number s within the range of [0, 1]; If s < q, directly take the target point of the unmanned aerial vehicle as the sampling point; Otherwise, a sampling point is randomly generated in the free space; a point Q is found in T1 and T2. samp The nearest node, which is the nearest neighbor node Q near , and record Q samp and Q near Euclidean distance between ; The formula used is as follows: ； ； Where, F obs and F free are the volume occupied by obstacles and the volume of free space, respectively.

4. A method for planning a path for a maritime UAV according to claim 2, characterized in that: In step S23, the calculation of the extended combined force specifically includes the following steps: Step S231: Calculate gravity; take the product of the Euclidean distance and the gravity gain coefficient as the target point pair Q near The gravitational pull of Step S232: Calculate repulsive force; for each obstacle, if the obstacle I a The center point and Q near The Euclidean distance between them is greater than the maximum influence range of the repulsive force, then I a Q near The repulsion is 0; otherwise, based on the Euclidean distance and repulsion gain coefficient, the product of the basic repulsion strength, target distance weight and direction vector is used as I a Q near Repulsion; move all obstacles to Q near The repulsive forces of are superimposed to obtain the total repulsive force; among them, I a is the ath obstacle; Step S233: Calculate the environmental force; according to the wind speed and direction of the sea breeze, combined with the environmental force gain coefficient, the effect of the ocean environment on Q is obtained. near environmental forces; Step S234: Synthesize the expansion force; synthesize the attraction, repulsion and environmental forces on Q near The impact of this will be expanded.

5. The method for planning a path for a maritime UAV according to claim 2, wherein: In step S27, the dual search tree connection is a pair of new Each node in another search tree is different, predicting the drone from Q new Xiang Q f The positions of all dynamic obstacles during the movement are detected when the drone moves from Q new Xiang Q f Whether it will collide with obstacles during movement; If there is a node Q f , so that the drone does not collide with obstacles, then splice Q new The path to the root node of the search tree to which it belongs and Q f The path to the root node of the search tree to which it belongs, and the candidate path J is obtained f ; Then from all candidate paths J f Select a shortest path as the initial path J first , complete the preliminary path planning; otherwise, go to step S22; where Q f Yes and Q new The fth node in another different search tree, f is the node index, J f Based on Q f Generated candidate paths.

6. The method for planning a path for a maritime UAV according to claim 1, wherein: In step S3, the path optimization is to optimize the preliminary path through pruning and path smoothing to obtain the final path; specifically, it includes the following steps: Step S31: Initialization; from the initial path J first Extract the ordered node sequence {Q1, Q2, ..., Q U }, and create an empty list E to store the optimized node sequence, initialize the current node index i=1; where U is the preliminary path J first The total number of nodes in Q1, Q2 and Q U The initial path J first The first, second and Uth nodes in Q1 and Q U They correspond to the starting point and target point of the drone respectively; Step S32: Evaluate and prune nodes segment by segment; includes the following steps: Step S321: Evaluate nodes segment by segment; Step S322: Pruning; for each node Q j , predicting drones from Q i Xiang Q j The positions of all dynamic obstacles during the movement are detected when the drone moves from Q i Xiang Q j Will it collide with obstacles during the movement? If there is a node Q j , so that the drone does not collide with obstacles, then Q j As a candidate node, the node with the smallest comprehensive evaluation value is selected from all candidate nodes as the optimal jump node Q best , Q best and Q i Add to E and update i to Q best Otherwise, only Q i Add to E; where Q i and Q j The initial path J first The i-th and j-th nodes in ; Step S323: Determine the simplified path; if i=U, connect the nodes in list E in order to obtain the simplified path; otherwise, if i=U-1, connect Q U Add it to E, and then connect the nodes in list E in order to obtain a simplified path; otherwise, add 1 to i and go to step S321 for the next iteration; Step S33: Path smoothing; perform cubic spline interpolation smoothing on the streamlined path to obtain the final path and complete path optimization.

7. The method for planning a path for a maritime UAV according to claim 6, wherein: In step S321, the segment-by-segment node evaluation is performed from the current node Q i Start and traverse all subsequent nodes Q j , j increases from i+1 to U, the energy consumption value is obtained by combining the air resistance energy consumption and the maneuvering energy consumption, the environmental threat value is obtained by combining the obstacle threat and the wind speed threat, and the weighted combination of the Euclidean distance, the energy consumption value and the environmental threat value is obtained. i to Q j The comprehensive evaluation value of this path.

8. The method for planning a path for a maritime UAV according to claim 1, wherein: In step S1, the definition of the three-dimensional grid map is to establish a three-dimensional coordinate system O-xyz with the starting point of the unmanned aerial vehicle as the origin, and collect data of the maritime unmanned aerial vehicle based on the three-dimensional coordinate system; The data of the maritime unmanned aerial vehicle includes marine environment data and the target point of the unmanned aerial vehicle; The marine environment data includes sea area spatial range data, obstacle data, and sea breeze data; Based on the sea area spatial range data, construct a three-dimensional grid map, mark static obstacles and dynamic obstacles in the three-dimensional grid map based on the obstacle data, and mark the starting point and target point of the unmanned aerial vehicle.

9. The method for planning a path for a maritime UAV according to claim 1, wherein: In step S4, the path navigation is to load the optimized final path into the navigation control system of the unmanned aerial vehicle. Under the guidance of the navigation control system, the unmanned aerial vehicle flies from the starting point to the target point according to the loaded path information.

Images