Marine ranching unmanned aerial vehicle and unmanned ship cooperative patrol path planning algorithm

Abstract

The application relates to the technical field of unmanned aerial vehicle and unmanned ship cooperation systems, in particular to a marine ranching unmanned aerial vehicle and unmanned ship cooperative patrol path planning algorithm, which comprises marine ranching environment gridding processing, unmanned ship path planning and smoothing processing and marine ranching patrol path planning. The marine ranching unmanned aerial vehicle and unmanned ship cooperative patrol path planning algorithm provided by the application is applicable to the overall environment of a marine ranching, provides endurance service for an unmanned aerial vehicle by taking advantages of flexible unmanned ship control, adaptability to harsh marine environments and the like, carries out coverage patrol of a marine ranching area by taking advantages of strong expandability, wide visual range and high patrol efficiency of the unmanned aerial vehicle, selects the optimal path in terms of time and energy consumption under the help of the unmanned ship and unmanned aerial vehicle cooperation system, and reduces patrol cost, and has important reference value for the patrol path planning of a marine ranching.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of unmanned aerial vehicle and unmanned ship cooperation system, more particularly to a marine ranching unmanned aerial vehicle and unmanned ship cooperative patrol path planning algorithm. BACKGROUND

[0002] As an important means of protecting fishery resources and restoring the ecological environment of the sea area, marine ranching is increasingly valued and has become a new sustainable fishery production model. However, due to the gaps and loopholes in the management of marine ranching, the phenomenon of stealing marine ranching is very serious. In order to reduce the phenomenon of stealing marine ranching, the best way is to arrange regular patrols and not to give the opportunity to the thieves. At present, for most fish farmers, the artificial patrol method is generally used, but the sea environment is harsh, and the patrol personnel will spend a lot of time and energy on the round trip, which is not conducive to the health of the patrol personnel. Therefore, it is urgent to adopt intelligent patrol means to replace the artificial patrol method to improve the patrol efficiency and reduce the patrol cost. At present, in order to realize intelligent patrol means, most of the unmanned aerial vehicles are used to cover and traverse the patrol of the marine ranching area, but the mileage of the unmanned aerial vehicle is limited by the battery energy, which cannot complete the task at one time, and needs to be charged in time. Inappropriate path will result in excessive energy consumption and time loss. SUMMARY

[0003] To achieve the above purpose, the technical scheme adopted by the present application is to provide a marine ranching unmanned aerial vehicle and unmanned ship cooperative patrol path planning algorithm, comprising the following steps:

[0004] Step 1: marine ranching environment gridding processing;

[0005] Step 2: unmanned ship path planning and smoothing processing;

[0006] Step 3: marine ranching patrol path planning.

[0007] Optionally, in step 1, the marine ranching area is mostly a convex polygon structure, and the marine ranching area to be patrolled is discretized into a plurality of target points, and a plurality of points are generated outside the marine ranching area as potential meeting points for the unmanned ship and the unmanned aerial vehicle to charge;

[0008] The gridding method of the convex polygon contour of the marine ranching is adopted to fill and expand the contour of the marine ranching, and binary processing is performed to construct the grid map of the marine ranching environment.

[0009] Optionally, the gridding method of the convex polygon contour of the marine ranching is:

[0010] First, according to the actual marine ranching boundary map, the vertex coordinates of the convex polygon area of the marine ranching are obtained, and the convex polygon contour of the marine ranching is drawn;

[0011] Then, the convex polygon contour is filled and expanded respectively to obtain a convex polygon region representing the marine ranching and a convex polygon contour region representing the potential convergence area, and the convex polygon region representing the marine ranching and the convex polygon contour region representing the potential convergence area are subjected to a binaryzation operation;

[0012] Secondly, the indexes of the marine ranching and the potential convergence point in the grid map are obtained;

[0013] Finally, the grid map of the marine ranching environment is constructed to complete the grid processing of the marine ranching environment.

[0014] Optionally, in step two, the unmanned ship travels in the sea area outside the marine ranching, an A* algorithm is used to plan the travel path of the unmanned ship, unnecessary inflection points in the A* planning path are removed, and a B-spline curve is used to smooth the path.

[0015] Optionally, the A* algorithm is used to plan the travel path of the unmanned ship, and the Euclidean distance is used as the cost estimation heuristic function of the A* algorithm, as shown in the following formula (1):

[0016] (1)

[0017] Wherein, D is the cost required to move one grid;

[0018] —current node coordinates;

[0019] —target node coordinates.

[0020] Optionally, the A* algorithm planning path contains unnecessary inflection points, and a B-spline curve is used to smooth the path, including the following steps:

[0021] Step 1: Obtain the basic path, initialize the traditional A* algorithm using the constructed grid map, and solve the basic path set Path0 containing unnecessary inflection points;

[0022] Step 2: Merge collinear points, judge whether the number of path points contained in Path0 is greater than 2, if yes, take the vector of the line connecting the last two points of Path0 as the reference, recursively judge whether there are collinear points, if yes, remove the middle redundant points and only keep the two end points. Obtain the path point set Path1;

[0023] Step 3: Remove unnecessary inflection points, connect the first and last path points of Path1 in turn, judge whether the connecting line segment is a passable domain, if yes, delete other points between the two points; if not, select new first and last path points in turn until all path points are traversed, and obtain the path point set Path2;

[0024] Step 4: Insert control points at the inflection points, determine whether the number of path points contained in Path2 is greater than 2, if yes, offset on the line segments on both sides of the inflection point to obtain two control points, and get path set Path3;

[0025] Step 5: B-spline curve interpolation smoothing, taking the path points of Path3 as the control points of the cubic B-spline, and the sampling points can be selected as total according to the path length to obtain the final path point set Path.

[0026] Optionally, the unmanned aerial vehicle performs B-spline curve smoothing on the path.

[0027] Optionally, in step three, the marine ranching patrol path planning includes single-block marine ranching patrol path planning.

[0028] Single-block marine ranching patrol path planning: the unmanned aerial vehicle uses a "bow-shaped" method to cover and traverse the single-block marine ranching area.

[0029] The specific steps of the single-block marine ranching area traversal algorithm are as follows:

[0030] Step S101: select a target ranching area, and execute step S102;

[0031] Step S102: there are MT potential merging points, select the i=0th potential merging point as the ranching patrol starting point, and execute step S103;

[0032] Step S103: determine whether all potential merging points have been used as patrol task starting points and solved, that is, determine whether i is less than MT, if yes, execute step S104; otherwise, execute step S112;

[0033] Step S104: set the i-th potential merging point as the patrol starting point, and initialize the related variables in the marine ranching area data structure Field, and execute step S105;

[0034] Step S105: determine whether all points in the marine ranching area have been visited, if yes, set i=i+1, and execute step S103; otherwise, execute step S106;

[0035] Step S106: according to the position of the take-off of the unmanned aerial vehicle, select a horizontal or vertical scanning mode, and execute step S107;

[0036] Step S107: perform a depth-first search on the marine ranching area, and execute step S108;

[0037] Step S108: backtrack the path obtained in step S107 to ensure that the unmanned aerial vehicle meets the mileage limit, and execute step S109;

[0038] Step S109: According to the last position of the UAV obtained in step S108, the nearest potential rendezvous point is selected, and step S110 is performed;

[0039] Step S110: The unmanned ship and the UAV go to the rendezvous point selected in step 8 respectively, the travel time of the UAV and the unmanned ship is calculated, and through backtracking operation, the unmanned ship is ensured to arrive at the rendezvous point before the UAV, and step S111 is performed;

[0040] Step S111: According to the situation of the visited patrol points, the unmanned ship is made to go to the nearest take-off point with the UAV, and step S112 is performed;

[0041] Step S112: The data structure Field of the solved single marine ranching field is saved, and step S113 is performed;

[0042] Step S113: The patrol path of the single marine ranching field is solved from different starting points, and the algorithm ends.

[0043] Optionally, in step S107, the exit conditions of the depth-first search include:

[0044] Condition 1: All target points have been visited or dead points have been encountered, and the exit is immediately exited;

[0045] Condition 2: When the remaining range of the UAV is negative, the exit is immediately exited;

[0046] Condition 3: When the remaining range of the UAV is less than the set safety range, the exit is exited when the UAV reaches the boundary point of the ranching field according to the scanning mode.

[0047] Optionally, in step three, the marine ranching path planning also includes multi-marine ranching path planning;

[0048] The multi-marine ranching path planning combines the single marine ranching path planning, converts the multi-marine ranching coverage problem into a TSP problem, and solves it by using an ant colony algorithm;

[0049] The solution result of the single marine ranching field includes: the flight distance of the UAV , the travel distance of the unmanned ship , the number of rendezvous , and the patrol time ;

[0050] The city set of the ant colony algorithm includes the patrol starting point and each potential rendezvous point. Different heuristic values are set for different solving targets;

[0051] The heuristic values include: the shortest total navigation distance of the unmanned ship, the least number of rendezvous charging, the shortest total time of the patrol task, and the addition of each target weight;

[0052] The total navigation distance of the unmanned ship is the shortest, as formula (2):

[0053] (2)

[0054] The number of times of meeting for charging is the least, as formula (3):

[0055] (3)

[0056] The total time of the patrol task is the shortest, as formula (4):

[0057] (4)

[0058] Add the weights of each target, as formula (5):

[0059] (5)

[0060] In formulas (2), (3), (4), and (5),

[0061] is the total number of pastures;

[0062] and are the distances of the unmanned ship to return to the starting point when starting and ending the traversal, respectively;

[0063] and are the corresponding time consumptions;

[0064] is the distance traveled by the unmanned ship when visiting the ith block of pasture;

[0065] is the number of times of meeting for the unmanned ship and the unmanned aerial vehicle when visiting the ith block of pasture;

[0066] is the time consumed when visiting the ith block of pasture;

[0067] is the weight of the distance of the unmanned ship;

[0068] is the weight of the number of times of meeting;

[0069] is the weight of the time consumed for the task.

[0070] The application is a kind of marine ranching unmanned aerial vehicle and unmanned ship cooperative patrol path planning algorithm, which is suitable for the global environment of marine ranching; the application provides endurance service for the unmanned aerial vehicle by using the advantages of unmanned ship, such as flexible operation and adaptability to harsh marine environment, covers the patrol of the marine ranching area by using the advantages of the unmanned aerial vehicle, such as strong expandability, wide field of view and high patrol efficiency; the unmanned ship and unmanned aerial vehicle cooperative system selects the optimal path in terms of time and energy consumption under the help of the algorithm, and reduces the patrol cost; the marine ranching patrol path planning has important reference value. BRIEF DESCRIPTION OF DRAWINGS

[0071] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0072] Figure 1 The flow chart of marine ranching environment gridding processing.

[0073] Figure 2 The definition diagram of marine ranching data structure.

[0074] Figure 3 The flow chart of single-block marine ranching patrol path planning area traversal algorithm. DETAILED DESCRIPTION

[0075] In order to make the technical problems, technical solutions and beneficial effects of the present application more clearly understood, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0076] The marine ranching unmanned aerial vehicle and unmanned ship cooperative patrol path planning algorithm provided by the embodiments of the present application will be described. A kind of marine ranching unmanned aerial vehicle and unmanned ship cooperative patrol path planning algorithm, comprising the following steps:

[0077] Step one, marine ranching environment gridding processing;

[0078] Step two, unmanned ship path planning and smoothing processing;

[0079] Step three, marine ranching patrol path planning. The marine ranching patrol path planning includes single-block marine ranching patrol path planning and multi-block marine ranching patrol path planning.

[0080] Specifically, in step one, the marine ranching area is mostly a convex polygon structure, and the marine ranching area to be patrolled is discretized into a plurality of target points. When working, it is considered that traversing all the target points is to complete the patrol of the area. In addition, a plurality of points are generated outside the marine ranching area as potential meeting points for the unmanned ship and the unmanned aerial vehicle to charge.

[0081] The grid method of the marine ranching convex polygon contour is adopted to fill and expand the marine ranching contour, and a binaryzation processing is performed to construct the grid map of the marine ranching environment.

[0082] Further, referring to Figure 1 , the grid method of the marine ranching convex polygon contour is as follows: first, according to the actual boundary map of the marine ranching, the vertex coordinates of the convex polygon area of the marine ranching are obtained to draw the convex polygon contour of the marine ranching; then, the filling and expansion operations are performed on the convex polygon contour to obtain the convex polygon area representing the marine ranching and the convex polygon contour area representing the potential meeting area, and the binaryzation operation is performed on the convex polygon area representing the marine ranching and the convex polygon contour area representing the potential meeting area; secondly, the indexes of the marine ranching and the potential meeting points in the grid map are obtained; finally, the grid map of the marine ranching environment is constructed to complete the grid processing of the marine ranching environment.

[0083] In step two, considering that the unmanned ship only drives in the sea area outside the marine ranching, on the basis of planning the driving path of the unmanned ship by the A* algorithm, unnecessary inflection points in the traditional A* planning path are removed, and a B-spline curve is used for smoothing the path. The specific algorithm for smoothing is as follows:

[0084] The smoothing algorithm can also be used for the path of the unmanned aerial vehicle.

[0085] Specifically, the A* algorithm is used to plan the driving path of the unmanned ship, and the Euclidean distance is used as the cost estimation heuristic function of the A* algorithm, as shown in the following formula (1):

[0086] (1)

[0087] wherein D is the cost required to move one grid;

[0088] —current node coordinates;

[0089] —target node coordinates.

[0090] The path planned by the traditional A* algorithm contains unnecessary inflection points, so the unnecessary inflection points need to be removed and the path needs to be smoothed.

[0091] Further, the algorithm for removing unnecessary inflection points and path smoothing includes the following steps:

[0092] Step 1: Obtain the basic path. Initialize the traditional A* algorithm with the constructed grid map to solve the basic path set Path0 containing unnecessary inflection points.

[0093] Step 2: Merge collinear points. Determine whether the number of path points in Path0 is greater than 2. If so, take the vector of the last two points in Path0 as the reference, recursively determine whether there are collinear points, and if so, remove the intermediate redundant points and keep only the two end points. Obtain the path point set Path1.

[0094] Step 3: Remove unnecessary inflection points. Connect the first and last path points of Path1 in turn, determine whether the connecting line segment is a passable domain, and if so, delete the other points between the two points; if not, select new first and last path points in turn until all path points are traversed. Obtain the path point set Path2.

[0095] Step 4: Insert control points at inflection points. Determine whether the number of path points in Path2 is greater than 2, and if so, offset by offset on both sides of the line segment at the inflection point to obtain two control points. Obtain the path set Path3.

[0096] Step 5: B-spline curve interpolation smoothing. Take the path points of Path3 as the control points of the cubic B-spline, and sample points can be selected as total according to the path length. Obtain the final path point set Path.

[0097] In step three, single-block marine ranching patrol path planning: since single-block marine ranching areas often have the characteristics of convex polygons, on the basis of the grid map, the unmanned aerial vehicle uses the "bow-shaped" method to cover and traverse the single-block marine ranching area.

[0098] The core of the area traversal algorithm includes depth-first search and exit condition judgment, selection of horizontal and vertical scanning, selection of merging points, and backtracking operation constrained by unmanned aerial vehicle energy and merging time.

[0099] The data structure Field of the single-block marine ranching area, as shown in Figure 2 , is used to store the planned path information.

[0100] The data structure of the monolithic ocean ranch includes muchang (an index set of ranch area), meet (an index set of potential meeting area), has_find_path (judging whether there is a solution starting from different starting points), uav_distances (total mileage of the UAV starting from different starting points), usv_distances (total mileage of the USV starting from different starting points), time_cost (total time cost of traversing the ranch starting from different starting points), uav_routes (a set of UAV paths starting from different starting points), usv_routes (a set of USV paths starting from different starting points), and charge_meets (a set of actual meeting points starting from different starting points).

[0101] The area traversal algorithm for the monolithic ocean ranch patrol path planning is shown in FIG. 1, and the specific steps of the area traversal algorithm are described as follows. Figure 3

[0102] Step S101: Start, select a target ranch area, and execute step S102.

[0103] Step S102: There are MT potential meeting points in total, select the ith (i = 0) potential meeting point as the starting point of the ranch patrol, and execute step S103.

[0104] Step S103: Judge whether all potential meeting points have been selected as the starting point of the patrol task and solved, i.e., judge whether i is less than MT, if yes, execute step S104; otherwise, execute step S112.

[0105] Step S104: Set the ith potential meeting point as the starting point of the patrol, and initialize the related variables in the ranch area data structure Field, and execute step S105.

[0106] Step S105: Judge whether all points in the ranch area have been visited, if yes, let i = i + 1, and execute step S103; otherwise, execute step S106.

[0107] Step S106: According to the position of the UAV takeoff, select the horizontal or vertical scanning mode.

[0108] Step S107: Execute the depth-first search (DFS) of the ranch area.

[0109] The exit conditions of the DFS include:

[0110] Condition 1: All target points have been visited or dead points are encountered, and the DFS is exited immediately;

[0111] Condition 2: The remaining mileage of the UAV is negative, and the DFS is exited immediately; ​

[0112] Condition 3: The remaining mileage of the UAV is less than the set safety mileage, and the UAV is exited according to the scanning mode when it reaches the pasture boundary point.

[0113] Step S108: A backtracking operation is performed on the path obtained in step S107 to ensure that the UAV meets the mileage limit.

[0114] Step S109: The last moment position of the UAV obtained in step S108 is selected as the nearest potential rendezvous point.

[0115] Step S110: The unmanned ship and the UAV respectively go to the rendezvous point selected in step S109, calculate the travel time of the UAV and the unmanned ship, and through a backtracking operation, ensure the time constraint, i.e. the unmanned ship reaches the rendezvous point before the UAV, and step S111 is executed.

[0116] Step S111: According to the situation of the visited patrol points, the unmanned ship takes the UAV to the nearest take-off point, and step S112 is executed.

[0117] Step S112: The data structure Field of the solved single marine pasture is saved, and step S113 is executed.

[0118] Step S113: The patrol path of the single marine pasture is solved from different starting points, and the algorithm ends.

[0119] Step three, multi-marine pasture patrol path planning:

[0120] Combined with the single marine pasture patrol path planning, i.e. on the basis of single marine pasture coverage patrol, the multi-marine pasture patrol coverage problem is converted into a TSP problem and solved by using an ant colony algorithm.

[0121] The solution result of the single marine pasture mainly includes: the unmanned ship travel distance , the number of rendezvous , and the patrol time .

[0122] The city set of the ant colony algorithm includes the patrol starting point and each potential rendezvous point. For different solving targets, different heuristic values can be set.

[0123] The heuristic value includes: the shortest total navigation distance of the unmanned ship, the least number of rendezvous charging, the shortest total time of the patrol task, and the addition of each target weight.

[0124] 1) The shortest total navigation distance of the unmanned ship, as formula (2):

[0125] (2)

[0126] 2) The least number of rendezvous charging, as formula (3):

[0127] (3)

[0128] 3) The total time of the patrol task is the shortest, as formula (4):

[0129] (4)

[0130] 4) Adding the weight of each target, as formula (5):

[0131] (5)

[0132] In the above formulas (2), (3), (4), (5),

[0133] is the total number of pastures;

[0134] and are the distances of the unmanned ship to return to the starting point when starting and ending the traversal, respectively;

[0135] and are the corresponding time losses;

[0136] is the distance traveled by the unmanned ship when visiting the ith block of pasture;

[0137] is the number of times the unmanned ship and the unmanned aerial vehicle meet when visiting the ith block of pasture;

[0138] is the time consumed when visiting the ith block of pasture;

[0139] is the unmanned ship distance weight;

[0140] is the meeting times weight;

[0141] is the task time consumption weight.

[0142] The present application is a kind of marine pasture unmanned aerial vehicle unmanned ship cooperative patrol path planning algorithm, applicable to marine pasture global environment;The present application utilizes the advantages of unmanned ship maneuvering flexibility, adaptability to harsh marine environment, etc. to provide endurance service for unmanned aerial vehicle, utilizes the advantages of strong expandability, wide field of view, high patrol efficiency, etc. to carry out the coverage patrol of marine pasture area;The unmanned ship and unmanned aerial vehicle cooperative system selects the path with optimal time and energy consumption under the help of algorithm, and reduces the patrol cost;The marine pasture patrol path planning has important reference value.

[0143] The above only describes preferred embodiments of the present application and is not used to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for collaborative patrol path planning between unmanned aerial vehicles (UAVs) and unmanned surface vessels (USVs) in marine ranching, characterized in that: Comprising the following steps: Step one, marine ranching environment gridding processing; Step two, unmanned ship path planning and smoothing processing; Step three, marine ranching patrol path planning; In step three, the marine ranching patrol path planning includes single block marine ranching patrol path planning; Single block marine ranching patrol path planning: the unmanned aerial vehicle adopts a "bow-shaped" mode to cover and traverse the single block marine ranching area; The specific steps of the single block marine ranching area traversal algorithm are as follows: Step S101: select the target ranching area and execute step S102; Step S102: there are MT potential meeting points, select the i=0th potential meeting point as the ranching patrol starting point and execute step S103; Step S103: determine whether all potential meeting points have been used as patrol task starting points and solved, that is, determine whether i is less than MT, if yes, execute step S104; Otherwise, execute step S112; Step S104: set the i-th potential meeting point as the patrol starting point and initialize the relevant variables in the marine ranching area data structure Field, and execute step S105; Step S105: determine whether all points in the marine ranching area have been visited, if yes, let i=i+1 and execute step S103; Otherwise, execute step S106; Step S106: select the horizontal or vertical scanning mode according to the position of the unmanned aerial vehicle take-off, and execute step S107; Step S107: perform a depth-first search on the marine ranching area, and execute step S108; Step S108: backtrack the path obtained in step S107 to ensure that the unmanned aerial vehicle meets the mileage limit, and execute step S109; Step S109: select the nearest potential meeting point according to the last position of the unmanned aerial vehicle obtained in step S108, and execute step S110; Step S110: the unmanned ship and the unmanned aerial vehicle go to the meeting point selected in step S109, calculate the travel time of the unmanned aerial vehicle and the unmanned ship, and through the backtracking operation, ensure that the unmanned ship reaches the meeting point before the unmanned aerial vehicle, and execute step S111; Step S111: according to the visited patrol points, let the unmanned ship take the unmanned aerial vehicle to the nearest take-off point, and execute step S112; Step S112: save the data structure Field of the solved single block marine ranching area, and execute step S113; Step S113: solve the patrol path of the single block marine ranching area from different starting points, and the algorithm ends; In step S107, the exit conditions of the depth-first search include: Condition 1: all target points have been visited or dead points have been encountered, immediately exit; Condition 2: the remaining mileage of the unmanned aerial vehicle is negative, immediately exit; Condition 3: the remaining mileage of the unmanned aerial vehicle is less than the set safety mileage, according to the scanning mode, exit when the unmanned aerial vehicle reaches the boundary point of the marine ranching area.

2. The method of claim 1, wherein: In step one, the marine ranching area is mostly a convex polygon structure, and the marine ranching area to be patrolled is discretized into target points, and a plurality of points outside the marine ranching area are generated as potential meeting points for the unmanned ship and the unmanned aerial vehicle to charge; The convex polygon contour of the marine ranching is rasterized to fill and expand the contour, and then binarized to construct a raster map of the marine ranching environment.

3. The method of claim 2, wherein: The rasterization method of the convex polygon contour of the marine ranching comprises the following steps: Firstly, the vertex coordinates of the convex polygon region of the marine ranching are obtained according to the actual boundary map of the marine ranching, and the convex polygon contour of the marine ranching is drawn; Then, the convex polygon contour is filled and expanded respectively to obtain the convex polygon region representing the marine ranching and the convex polygon contour region representing the potential merging region, and the convex polygon region representing the marine ranching and the convex polygon contour region representing the potential merging region are binarized; Secondly, the indices of the marine ranching and the potential merging points in the raster map are obtained; Finally, the raster map of the marine ranching environment is constructed to complete the rasterization of the marine ranching environment.

4. The method of claim 1, wherein: In step two, the unmanned ship travels in the sea area outside the marine ranching area, adopts A An algorithm is used to plan the travel path of the unmanned ship, remove A unnecessary inflection points in the planned path, and use B-spline curve to smooth the path.

5. The method of claim 4, wherein: A The algorithm plans the travel path of the unmanned ship, and adopts Euclidean distance as A The cost estimation heuristic function of the algorithm is shown in the following formula (1): （1） Wherein, D is the cost required to move one grid; - current node coordinates; - Target node coordinates.

6. The method of claim 5, wherein: Remove A The algorithmic planned path contains unnecessary inflection points and is smoothed using B-spline curves, including the following steps: Step 1: Get the base path, initialize the traditional A algorithm with the constructed grid map, and solve the base path set Path0 containing the non-essential inflection points; Step 2: Merge collinear points, judge whether the number of path points contained in Path0 is greater than 2, if yes, take the vector of the line connecting the last two points of Path0 as the reference, recursively judge whether there are collinear points, if yes, remove the middle redundant points and only keep the two end points to obtain the path point set Path1; Step 3: Remove unnecessary inflection points, connect the first and last path points of Path1 in turn, judge whether the connecting line segment is a passable domain, if yes, delete other points between the two points; if not, select new first and last path points in turn until all path points are traversed to obtain the path point set Path2; Step 4: Insert control points at inflection points, judge whether the number of path points contained in Path2 is greater than 2, if yes, offset offset on the line segments on both sides of the inflection point to obtain two control points to obtain the path set Path3; Step 5: B-spline curve interpolation smoothing, taking the path points of Path3 as the control points of the cubic B-spline, the sampling points can be selected as total according to the path length to obtain the final path point set Path.

7. The method of claim 4, wherein: The UAV uses B-spline curve to smooth the path.

8. The method of claim 1, wherein: In step three, the marine ranching patrol path planning further comprises multi-block marine ranching patrol path planning; The multi-block marine ranching patrol path planning combines the single-block marine ranching patrol path planning, converts the multi-ranching patrol coverage problem into a TSP problem, and solves it by using an ant colony algorithm; The monolithic mariculture solution results include: UAV flight distance , unmanned ship travel distance , number of mergers , and patrol time ; The city set of the ant colony algorithm comprises a patrol starting point and each potential merging point, and different heuristic values are set for different solving targets; The heuristic values comprise: shortest total navigation distance of the unmanned ship, least number of merging and charging times, shortest total time of the patrol task, and addition of weights of each target; The shortest total navigation distance of the unmanned ship is as formula (2): （2） The least number of merging and charging times is as formula (3): （3） The shortest total time of the patrol task is as formula (4): （4） The addition of weights of each target is as formula (5): （5） In formulas (2), (3), (4) and (5), Total Pasture = Total Pasture Number; and are the distances from the start point for the unmanned ship to begin and end the traversal, respectively; and is the corresponding time lag; a distance travelled by the unmanned ship to access the ith paddock; the number of rendezvous of the unmanned ship and the unmanned aerial vehicle for accessing the i-th pasture; time spent in accessing the ith paddock; Unmanned ship mileage weight; - number of mergings weight; — task time consuming weight.

Images