Unmanned aerial vehicle cruise path planning method for multi-priority dynamic coverage

Abstract

The invention discloses an unmanned aerial vehicle cruise path planning method for multi-priority dynamic coverage, and the method comprises the steps: carrying out the sorting of the detection priorities of different parts of a region according to the traversal period constraint, carrying out the high-frequency cruise detection of key parts, and carrying out the low-frequency cruise detection of general parts; secondly, according to the priority ranking, recursively constructing a series of cyclic cruise path planning schemes by adopting a stochastic optimization algorithm, so that the region is dynamically covered and meets corresponding traversal period constraints; and finally, analyzing and evaluating the cruise path planning scheme from two aspects of redundancy and priority, and outputting an optimal planning scheme. Compared with a traditional path planning method, cyclic cruise detection is carried out according to the priority, the whole area can be dynamically covered with emphasis, the cooperative cruise detection efficiency of the unmanned aerial vehicle is effectively improved, and the area change is accurately mastered; in addition, the cooperative cruise path can be adaptively adjusted according to the priority change, unmanned detection resources are efficiently utilized, and the method flexibly adapts to the dynamic coverage demand change.

Description

TECHNICAL FIELD

[0001] The application relates to a UAV cruise path planning method for multi-priority dynamic coverage. BACKGROUND

[0002] A traditional UAV cooperative detection path planning method for static area coverage plans an optimal navigation route capable of traversing the coverage area under the premise of meeting certain (some) performance indicators, avoiding obstacles and threats, and the traversal frequency of the entire area is usually 1. For example, Berger et al. assigned scanning tasks to each UAV based on the capability difference between the platform and the sensor, and achieved effective scan trajectory planning (BERGER C, WZOREK M, KVARNSTRÖMJ, et al. Area coverage with heterogeneous UAVs using scan patterns [C],2016IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR), Lausanne, Switzerland, 2016); Wang Zili et al. proposed an effective planning algorithm for assisting UAVs to autonomously complete the given area coverage search (Wang Zili, Luo Delin, Wu Shunxiang, Concave polygon area coverage UAV path planning method [J],Aircraft weapons, 2019, 26(1): 95-100); Zhang Xiaomeng et al. proposed an improved multi-UAV coverage path planning method based on region decomposition, Z-shaped scanning mode and shape turning method (Zhang Xiaomeng, Hu Yongjiang, Li Wenguang, et al. An improved multi-UAV coverage path planning method [J],Journal of Weapon Equipment Engineering, 2020, 41(10): 21-221). However, due to the dynamic complexity of the region, such as the movement of the detection target and the spread of the fire, it is difficult to grasp the dynamic changes of the region, especially the development and change of the key parts, through single traversal. SUMMARY

[0003] The application provides a multi-priority dynamic coverage UAV cruise path planning method that can flexibly and efficiently grasp the dynamic changes of the region by cyclically cruising and detecting different parts of the region.

[0004] Technical scheme: To solve the above problems, the application adopts a multi-priority dynamic coverage UAV cruise path planning method, which comprises the following steps:

[0005] (1) grid division is performed on the cruise area to obtain a plurality of cruise grids, the cruise grids are numbered and the traversal period of each cruise grid is obtained;

[0006] (2) the cruise grids are sorted according to the traversal period from short to long, and when the traversal period is the same, the cruise grids are sorted according to the number from small to large, so as to obtain the priority order of the cruise grids;

[0007] (3) on the basis of the current cruise path planning scheme, the cruise grid with the highest priority and the empty set of traversal sets is taken as the first priority coverage grid, and the cruise path of the priority coverage grid is planned: other cruise grids with a distance from the priority coverage grid within half of the traversal period are taken as candidate cruise grids, and a plurality of candidate cruise paths capable of covering the priority coverage grid are planned according to the priority coverage grid and the candidate cruise grids;

[0008] (4) on the basis of a plurality of candidate cruise paths, depth-first recursion is performed according to the priority order, and the subsequent priority coverage grid and a plurality of candidate cruise paths covering the grid are recursively explored, so as to find a cruise path planning scheme capable of covering all cruise grids;

[0009] (5) all cruise path planning schemes are evaluated, and the optimal cruise path planning scheme is output, and dynamic periodic cruise is performed according to the optimal cruise path planning scheme.

[0010] Further, the cruise area includes a no-fly zone and a cruise area, and the cruise area is divided into cruise grids after grid division.

[0011] Further, the higher the cruise priority of the cruise grid is, the shorter the corresponding traversal period is, and the traversal period is an even number of time steps, and in each time step, the navigation distance of each unmanned aerial vehicle is one cruise grid.

[0012] Further, for the cruise area after grid division, a shortest path distance matrix and a shortest path set matrix are constructed. and the shortest path set matrix , other cruise grids with a distance from the priority coverage grid within half of the traversal period are taken as candidate cruise grids;

[0013] Each element in the shortest path distance matrix represents the shortest cruise path length between cruise grid and cruise grid .

[0014] Each element The elements in Indicates the cruise grid and Cruise Grid The set of shortest paths between .

[0015] Furthermore, in step (3), when planning several candidate cruise paths in the candidate cruise grid, the first The optimal candidate cruise paths that prioritize covering grids, where To explore the threshold;

[0016] The optimization objective function and constraints for selecting candidate cruise paths are:

[0017]

[0018]

[0019] in, To be The set of cruise grids surrounded by candidate cruise paths, Passing the cruise grid The set of cruise path numbers, is the ceiling function, For cruise grid The traversal cycle, Prioritize coverage squares The traversal cycle, For the The number of the first cruise square that the candidate cruise path passes through, For the The candidate cruise path passes through The number of the cruise grid, Prioritize coverage squares Number, For the candidate cruise paths, For the The length of the candidate cruise paths.

[0020] Furthermore, in step (3), a random optimization algorithm is used when planning a number of candidate cruise paths within the candidate cruise grid.

[0021] Furthermore, the scheme evaluation of all cruise path planning schemes includes redundancy evaluation and priority evaluation, and the calculation formula of the redundancy evaluation is:

[0022]

[0023] in, Plan a cruise route a redundancy score of the cruise path planning scheme, a path length of the cruise path, a total number of the cruise squares;

[0024] The calculation formula of the priority evaluation is:

[0025]

[0026] wherein, a priority score of the cruise path planning scheme a traversal period of the cruise square The lower the score is, the higher the priority is.

[0027] Further, the cruise path planning scheme with the lowest redundancy score is selected as the optimal cruise path planning scheme, and if there are multiple cruise path planning schemes with the lowest redundancy score, the cruise path planning scheme with the lowest priority score is selected as the optimal cruise path planning scheme from the cruise path planning schemes with the lowest redundancy score, and the number of the cruise paths in the optimal cruise path planning scheme is the number of the unmanned aerial vehicles required for cruising.

[0028] The application also adopts a computer device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the above method when executing the computer program.

[0029] The application also adopts a computer readable storage medium, which stores a computer program, and the computer program implements the steps of the above method when executed by a processor.

[0030] Advantages: Compared with the prior art, the application has the following advantages: (1) the cyclic cruise detection is performed according to the priority, the whole area can be dynamically covered with emphasis, the unmanned aerial vehicle cooperative cruise detection efficiency is effectively improved, and the area change is accurately mastered; (2) the cooperative cruise path can be adaptively adjusted according to the priority change, the unmanned detection resources are efficiently utilized, and the dynamic coverage demand change is flexibly adapted. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 a flowchart of the cruise path planning method of the application;

[0032] Figure 2 an unmanned aerial vehicle cooperative cruise area in the embodiment of the application;

[0033] Figure 3 a candidate cruise area of the square in the embodiment of the application;

[0034] Figure 4 ​​This is the optimal cruise path planning solution output by the implementation case in this invention. DETAILED DESCRIPTION

[0035] like Figure 1 As shown in the figure, in this embodiment, a method for UAV cruise path planning for multi-priority dynamic coverage is provided. The UAV cooperative cruise area includes two parts: a no-fly zone and a cruise zone. The cruise zone is divided into two parts: a no-fly zone and a cruise zone. It is composed of interconnected squares, each of which is a cruising square, which are recorded as 、 、…、 ; Square ( ) is recorded as The higher the priority level, the shorter the corresponding traversal cycle, and the lower the priority level, the longer the corresponding traversal cycle. The traversal cycle value is an even number of time steps. In each time step, the navigation distance of each drone is 1 square, and the flight direction is Axis or Axis; UAV cruise path planning for multi-priority dynamic coverage. The UAV cruise path is planned so that each grid can be visited by the UAV according to its traversal cycle, and the dynamic changes of the area can be grasped more flexibly and accurately. The specific steps for planning the UAV cruise path are as follows:

[0036] Step 1: Model the shortest path in the cruise area after grid division: Construct the shortest path distance matrix and the shortest path set matrix , specifically including:

[0037] Step 1.1, shortest path distance matrix construction: Construction dimensional shortest path distance matrix , each element in the shortest path distance matrix Represents square pairs The shortest cruise path length between 、 , dynamic programming method is used for calculation (for dynamic programming method, see Shi Shaojian, Research on Dynamic Programming Algorithm [J], Computer Knowledge and Technology, 2020, (18): 48-49, which will not be repeated in this embodiment), assuming that the grid For square Neighbors, , then the square pair The shortest path length between them is equal to the square and squares Shortest path length and squares and squares Shortest path length The minimum value of the sum is as shown below;

[0038] (1)

[0039] Step 1.2, shortest path set matrix construction: Construction dimensional shortest path set matrix , each element in the shortest path set matrix Represents square pairs The set of shortest paths between , , because the same shortest path length may correspond to multiple shortest paths, so , also use the dynamic programming method to find the neighboring squares that the shortest path passes through , thus obtaining the grid pair The shortest path between is:

[0040] (2)

[0041] Step 2: Cruise path planning initialization: Cruise path planning initialization includes global element initialization such as cruise path exploration threshold setting, grid coverage priority sorting, and planning scheme set initialization. It also includes recursive element initialization such as grid index initialization, grid coverage matrix initialization, and current planning scheme initialization. Specifically, it includes:

[0042] Step 2.1. Cruise path exploration threshold setting: Set the exploration threshold , which means that for each square, only the previous squares that pass through the square according to its traversal cycle are explored. Optimal cruising paths;

[0043] Step 2.2, grid coverage priority sorting: The so-called grid coverage means that the cruise path loops through the grid, the loop period is less than or equal to the traversal period of the grid, and there are no untraversed grids in the area surrounded by the cruise path; priority sorting means sorting the grids by traversal period, and giving priority to covering the grids with short traversal periods when constructing the cruise path; specifically, the grids to be traversed are sorted. The squares are sorted from short to long according to the traversal period. If the traversal period is the same, they are sorted from small to large according to the square number, thus obtaining the square coverage priority sorting. ,satisfy , and subsequently, the squares with short traversal periods are covered first according to this sorting;

[0044] Step 2.3, initialization of planning scheme set: constructing planning scheme set , solution set Each element in is a cruise path planning scheme, and each cruise path planning scheme is a set consisting of several cruise paths that meet the grid coverage requirements. The scheme set Initialized to an empty set, that is ;

[0045] Step 2.4, grid index initialization: create grid index , indicating the priority order in grid coverage The starting exploration index in is initialized to 1 because the squares with short traversal periods are explored first;

[0046] Step 2.5, grid coverage matrix initialization: construct dimensional grid covering matrix ,matrix Each element in For the reason of passing through the grid The set of cruise path numbers, 1, 2, ..., , the cruise path number is the number of the grid that the path covers first, and the grid coverage matrix Each element in is initialized to an empty set, that is, ;

[0047] Step 2.6, current planning scheme initialization: current cruise path planning scheme Said that the plan is a set consisting of several cruise paths that meet the grid coverage requirements, initialized to an empty set. ;

[0048] Step 3: Recursive exploration of planning schemes: Current tour path planning scheme Based on the grid index Corresponding square First, according to the shortest path distance matrix and the shortest path set matrix , perform depth-first recursive exploration according to priority, and explore the front of the grid that covers the grid. The optimal cruise path is recursively explored, and the recursive exploration process of the planning scheme is recorded as , the specific sub-steps include 3.1-3.3;

[0049] Step 3.1, candidate cruise area positioning: number Square As the priority coverage square, obtain its traversal cycle , according to the shortest path distance matrix and the shortest path set matrix , locate the distance grid exist All squares within are squares Candidate cruising areas ;

[0050] Step 3.2, candidate cruise path mining: using random optimization algorithm, in the candidate cruise area Before excavation Priority Covering Squares The optimal candidate cruise path, , , the constraints that the candidate cruise path must meet include: (1) The candidate cruise path starts from the priority coverage grid Start by going to the priority coverage square End, that is , (2) The length of the candidate cruise path is square The traversal cycle ,Right now , (3) Assume that the set of squares enclosed by the candidate cruise path is denoted as , then the set All squares in are on the candidate path, that is, , , the newly added square is a set In the path Cover the cruise grid that has not been covered by other cruise paths before. The optimization goal is to maximize the weighted sum of the traversal cycles of the newly added covered grid. ,Right now

[0051]

[0052] (3)

[0053] Since the selected priority coverage squares are sorted from short to long according to the traversal period, all the squares passed by the candidate cruise path are either covered by the previous cruise path or covered by the current cruise path, and the cycle period can meet the traversal period constraint requirements of the square;

[0054] Step 3.3, candidate cruise path recursion: for each front The optimal candidate cruise paths, , , perform the following sub-steps 3.3.1-3.3.5 respectively:

[0055] Step 3.3.1. Update the current planning scheme: Update the current patrol path planning scheme and add candidate patrol paths ,Right now ;

[0056] Step 3.3.2, square coverage matrix update: for the candidate cruise path After all squares, update its traversal set in the square coverage matrix, add the number of priority coverage squares in the traversal set ; That is , , ;

[0057] Step 3.3.3, square index update: update the square index, update the index corresponding to the square whose next traversal set is empty, denoted as , satisfying and When all squares corresponding to the traversal set are empty, update the square index to ;

[0058] Step 3.3.4, square index judgment: if the square index satisfies , it indicates that all squares have been covered by the cruise path, and there is no need to perform the next level recursion, update the cruise path planning scheme set , end the current level of candidate cruise path recursion process, return to the previous step, otherwise continue to the next step;

[0059] Step 3.3.5, next level exploration: continue to recursively mine candidate cruise paths at the next level ; that is, for each cruise path planning scheme , return to step 3 to continue the next level cruise path mining.

[0060] Step 4, planning scheme judgment: judge whether the planning scheme set is empty, if so, output "no cruise planning scheme found", end the process, otherwise, continue to step 5;

[0061] Step 5, planning scheme evaluation: each planning scheme in the planning scheme set is a set composed of several cruise paths, and each cruise path in the scheme covers several squares, redundancy and priority evaluation are performed for each planning scheme, sub-steps include 5.1, 5.2:

[0062] Step 5.1, redundancy evaluation: calculate the ratio of the sum of the lengths of all cruise paths in each planning scheme to the total number of squares , as the redundancy score, as shown in the following formula:

[0063] (4)

[0064] The lower the score, the lower the redundancy of the covered grid;

[0065] Step 5.2, priority evaluation: calculating the priority score of each planning scheme The average traversal period of the grid covered by the inner cruise path is taken as the priority score, as shown in the following formula:

[0066] (5)

[0067] The lower the score, the shorter the average traversal period of the covered grid, and the higher the priority;

[0068] Step 6, output of the optimal planning scheme: if there is only one planning scheme with the lowest redundancy, select it as the optimal scheme, as shown in the following formula:

[0069] (6)

[0070] If there is more than one planning scheme with the lowest redundancy, select the planning scheme set with the lowest redundancy with the highest priority as the optimal planning scheme, as shown in the following formula:

[0071] (7)

[0072] Calculate the number of cruise paths of the optimal planning scheme , that is, the number of unmanned aerial vehicles required for cruising, as shown in the following formula:

[0073] (8)

[0074] Output the optimal planning scheme and the number of unmanned aerial vehicles required for cruising .

[0075] The unmanned aerial vehicle cruise path planning method for multi-priority dynamic coverage proposed in the present application organizes unmanned aerial vehicles to cooperatively cruise and detect the region, and the key parts are cyclically cruised and detected according to a short traversal period, and the general parts are cyclically cruised and detected according to a long traversal period, so that the unmanned aerial vehicle resources can be efficiently utilized, and the dynamic changes of the region can be flexibly mastered.

[0076] In this embodiment, 8 grids are taken as examples, each grid is a cruise grid, and the unmanned aerial vehicle cooperative cruise region includes a no-fly zone and a cruise zone, wherein the cruise zone is composed of 8 connected grids, which are respectively denoted as , , , Figure 2 as shown in the following figure, the grid is connected to Both are important parts with high priority, and their traversal period is 2 time steps, and the square and Both are parts with medium priority, and their traversal period is 4 time steps, and the remaining four squares , , , are parts with general priority, and their traversal period is 6 time steps. In each time step, the flight distance of each UAV is 1 square, and the flight direction is axis or axis. The UAV cruise path planning for multi-priority dynamic coverage completes the UAV cruise path planning, so that each square can be visited by the UAV in a cycle according to its traversal period, and the regional dynamic change can be more flexibly and accurately mastered. The UAV cruise path planning steps are as follows:

[0077] Step 1.1, shortest path distance matrix construction: construct a dimensional shortest path distance matrix , each element in the shortest path distance matrix represents the shortest cruise path length between the square pair , wherein , , and the dynamic programming method is used for calculation. The matrix construction result is shown in the following table:

[0078] Table 1 shortest path distance matrix construct

[0079]

[0080] Step 1.2, shortest path set matrix construction: construct a dimensional shortest path set matrix , each element in the shortest path set matrix represents the shortest path set between the square pair , wherein , , and the shortest path set matrix is shown in Table 2:

[0081] Table 2 shortest path set matrix construct

[0082]

[0083] Step 2, cruise path planning initialization: cruise path planning initialization includes global element initialization such as cruise path exploration threshold setting, grid coverage priority sorting and planning scheme set initialization, and also includes recursive element initialization such as grid index initialization, grid coverage matrix initialization and current planning scheme initialization, and its sub-steps include 2.1-2.6;

[0084] Step 2.1, cruise path exploration threshold setting: set the exploration threshold , which means that for each grid, only the first 2 optimal cruise paths that pass through the grid in its traversal period are explored;

[0085] Step 2.2, grid coverage priority sorting: sort the grid to be traversed according to the traversal period from short to long, if the traversal period is the same, sort by grid number from small to large, thus obtaining the grid coverage priority sorting ;

[0086] Step 2.3, planning scheme set initialization: construct the planning scheme set , each element in the scheme set is a cruise path planning scheme, and each cruise path planning scheme is a set composed of several cruise paths that meet the grid coverage requirements, and the scheme set is initialized as an empty set, i.e. ;

[0087] Step 2.4, grid index initialization: establish the grid index , which represents the starting exploration index in the grid coverage priority sorting , because the grid with short traversal period is preferentially traversed, its value is initialized to 1, i.e. ;

[0088] Step 2.5, grid coverage matrix initialization: construct an 8-dimensional grid coverage matrix , each element in the matrix is a set composed of cruise path numbers that pass through the grid , 1, 2, …, 8, the cruise path number is the number of the path that preferentially covers the grid, and each element in the grid coverage matrix is initialized as an empty set, i.e. ;

[0089] Step 2.6, current planning scheme initialization: the current cruise path planning scheme is represented, and the scheme is a set composed of several cruise paths that meet the grid coverage requirements, and is initialized as an empty set, ;

[0090] Step 3: Recursive exploration of planning schemes:

[0091] Step 3 （1） , First-level candidate cruise path mining: Current tour path planning scheme Based on the grid index Corresponding square First, according to the shortest path distance matrix and the shortest path set matrix , explore the previous grid as needed The optimal cruise path is recursively explored, and the recursive exploration process of the planning scheme is recorded as , specifically including the following;

[0092] Step 3.1 （1） , Candidate cruise area positioning: the number Square As the priority coverage square, obtain its traversal cycle , according to the shortest path distance matrix and the shortest path set matrix , locate the distance grid exist All squares within are squares Candidate cruising areas ,like Figure 3 As shown;

[0093] Step 3.2 （1） , Candidate cruise path mining: Using random optimization algorithm, in the candidate cruise area An optimal candidate cruise path is mined. = , the constraints of the candidate cruise path include: (1) the candidate cruise path starts from the priority coverage grid Start by going to the priority coverage square End, (2) The length of the candidate cruise path is square The traversal cycle 2, (3) the grid set traversed by the candidate cruise path , then all the squares in the area are on the candidate path, and the weighted sum of the traversal cycles of the newly added covered squares is ;

[0094] Step 3.3 （1） , Candidate cruise path recursion: for the optimal candidate cruise path = , recursively execute the following substeps:

[0095] Step 3.3.1 （1）, Current planning scheme update: update the current tour path planning scheme, add candidate cruise path , i.e. ;

[0096] Step 3.3.2 （1） , Grid coverage matrix update: for the candidate cruise path , update its traversal set in the grid coverage matrix, and add the priority coverage No. 1, as shown in Table 3:

[0097] Table 3 Grid coverage matrix

[0098]

[0099] Step 3.3.3 （1） , Grid index update: update the grid index, update the index corresponding to the grid whose next traversal set is empty, denoted as ;

[0100] Step 3.3.4 （1） , Grid index judgment: if the grid index does not satisfy , continue to the next step;

[0101] Step 3.3.5 （1） , Next level exploration: continue to explore the candidate cruise path in the second level ;

[0102] Step 3 （2） , Candidate cruise path recursion (L2): based on the current tour path planning scheme , start from the grid Corresponding to the grid , according to the shortest path distance matrix And the shortest path set matrix , use a random optimization algorithm to mine two optimal candidate cruise paths, for the first two optimal candidate cruise paths, And , the traversal period weighted sum of the newly added coverage grids of the two candidate paths is:

[0103]

[0104]

[0105] For the first two optimal candidate cruise paths, And , respectively recursively execute the following steps:

[0106] Step 3-1（2） :against , update the current planning scheme to , update the grid covering matrix , for candidate cruise paths After passing through all squares (2, 5, 8), update its traversal set in the square coverage matrix and add priority coverage to the traversal set No. 2, as shown in Table 4 below:

[0107] Table 4 Grid Covering Matrix

[0108]

[0109] Update the grid index to , continue to mine candidate cruise paths at the third level ;

[0110] Step 3-1 （3） , Candidate cruise path recursion (L3): Current tour path planning scheme Based on the grid index Corresponding square First, according to the shortest path distance matrix and the shortest path set matrix , use random optimization algorithm to dig out an optimal candidate cruise path, and give priority to covering the grid , , perform the following steps respectively:

[0111] against , update the current planning scheme to , update the grid covering matrix , for candidate cruise paths After passing through all squares (3, 4, 5, 6, 7, 8), update its traversal set in the square coverage matrix and add priority coverage to the traversal set No. 4, as shown in Table 5 below:

[0112] Table 5 Grid Covering Matrix

[0113]

[0114] When all grid corresponding traversal sets are not empty, update the grid index , determine if the grid index satisfies , indicating that all squares have been covered by the cruise path, there is no need to recurse to the next level, and update the cruise path planning solution set , return to level 2;

[0115] Step 3-2（2） :For another optimal candidate cruise path at the second level , continue to target , update the current planning scheme to , update the grid covering matrix , for candidate cruise paths After passing through all squares (2, 5, 4), update its traversal set in the square coverage matrix and add priority coverage to the traversal set No. 2, as shown in Table 6 below,

[0116] Table 6 Grid Covering Matrix

[0117]

[0118] Update the grid index to , continue to mine candidate cruise paths at the third level ;

[0119] Step 3-2 （3） , For the third level candidate cruise path recursion (L3) of another optimal candidate cruise path in the second level: Current cruise path planning scheme Based on the grid index Corresponding square First, according to the shortest path distance matrix and the shortest path set matrix , use random optimization algorithm to dig out an optimal candidate cruise path, and give priority to covering the grid , , perform the following steps respectively;

[0120] against , update the current planning scheme to , update the grid covering matrix , for candidate cruise paths After passing through all squares (3, 4, 5, 6, 7, 8), update its traversal set in the square coverage matrix and add priority coverage to the traversal set No. 8, as shown in Table 7 below:

[0121] Table 7 Grid Covering Matrix

[0122]

[0123] When all grid corresponding traversal sets are not empty, update the grid index , determine if the grid index satisfies , indicating that all the squares have been covered by the cruise path, and there is no need to perform the next level recursion, updating the cruise path planning scheme set , return to the 2nd level, the 1st level in turn;

[0124] Step 4, planning scheme judgment: judging whether the planning scheme set is empty, continue to step 5;

[0125] Step 5, planning scheme evaluation: redundancy and priority evaluation are respectively performed on two planning schemes in the planning scheme set :

[0126] Step 5.1, redundancy evaluation: the ratio of the sum of the lengths of all cruise paths in each planning scheme to the total number of squares is taken as the redundancy score, and the redundancy scores of the two planning schemes are the same, both of which are :

[0127] Step 5.2, priority evaluation: the average traversal period of the squares covered by the cruise path in each planning scheme is calculated, the priority score of the 1st planning scheme is , and the priority score of the 2nd planning scheme is , so the score of the 1st scheme is the lowest, and its priority is higher;

[0128] Step 6, optimal planning scheme output: the redundancy scores of the two planning schemes are the same, but the priority evaluation score of the 1st planning scheme is lower, so the optimal planning scheme is: , the number of cruise paths of the optimal planning scheme is calculated, that is, the number of drones required for cruising, the optimal planning scheme and the number of drones required for cruising are output, as shown in Figure 4 , wherein the three cruise paths , and are represented by blue, brown and green respectively;

[0129] In the present application, the unmanned aerial vehicle cooperative cruising area is modeled by a plurality of connected squares, and the cruising speed of the unmanned aerial vehicle is uniformly assumed to be one square per time step. In actual situations, other models such as Venn diagram can be flexibly used to represent more complex cruise units, and different values of the cruising speed of the unmanned aerial vehicle can also be set in combination with different specific scene requirements.

Claims

1. A method for UAV cruise path planning with multi-priority dynamic coverage, characterized in that: The following steps are involved: (1) Divide the cruise area into grids to obtain a number of cruise grids, number the cruise grids and obtain the traversal period of each cruise grid; (2) Sort the cruise grids from short to long according to the traversal period. If the traversal periods are the same, sort them from small to large according to the numbers of the cruise grids, so as to obtain the priority order of the cruise grids. (3) Based on the current cruise path planning scheme, the cruise grid with the highest priority and an empty traversal set is used as the first priority coverage grid, and a cruise path is planned for the priority coverage grid: other cruise grids within half a traversal period from the priority coverage grid are obtained as candidate cruise grids, and several candidate cruise paths that can cover the priority coverage grid are planned based on the priority coverage grid and the candidate cruise grids; (4) Based on several candidate cruise paths, perform depth-first recursion according to priority sorting, recursively explore the subsequent priority covered grid and several candidate cruise paths covering the grid, and find a cruise path planning scheme that can cover all cruise grids; (5) Evaluate all cruise path planning schemes, output the optimal cruise path planning scheme, and perform dynamic periodic cruise according to the optimal cruise path planning scheme.

2. The method for UAV cruise path planning with multi-priority dynamic coverage according to claim 1 is characterized in that: The cruising area includes a no-fly zone and a cruising area, and the cruising area is divided into grids to obtain cruising grids.

3. The method for UAV cruise path planning with multi-priority dynamic coverage according to claim 1 is characterized in that: The traversal cycle is an even number of time steps, and within each time step, the navigation distance of each drone is one cruising square.

4. The method for UAV cruise path planning with multi-priority dynamic coverage according to claim 1 is characterized in that: Construct the shortest path distance matrix for the cruise area after grid division and the shortest path set matrix In step (3), according to the shortest path distance matrix and the shortest path set matrix , obtaining other cruising squares whose distance from the priority coverage square is within half a traversal period as candidate cruising squares; The shortest path distance matrix The elements in Indicates the cruise grid and Cruise Grid The shortest cruise path length between The shortest path set matrix The elements in Indicates the cruise grid and Cruise Grid The set of shortest paths between .

5. The method for UAV cruise path planning with multi-priority dynamic coverage according to claim 1 is characterized in that: In the step (3), when planning several candidate cruise paths in the candidate cruise grid, the first The optimal candidate cruise paths that prioritize covering grids, where To explore the threshold; The optimization objective function and constraints for selecting candidate cruise paths are: ； ； in, For the The set of cruise grids surrounded by candidate cruise paths, Passing the cruise grid The set of cruise path numbers, is the ceiling function, For the cruise grid The traversal cycle, Prioritize coverage squares The traversal cycle, For the The number of the first cruise square that the candidate cruise path passes through, For the The candidate cruise path passes through The number of the cruising grid, Prioritize coverage squares Number, For the candidate cruise paths, For the The length of the candidate cruise paths.

6. The method for UAV cruise path planning with multi-priority dynamic coverage according to claim 5 is characterized in that: In the step (3), a random optimization algorithm is used when planning a number of candidate cruise paths within the candidate cruise grid.

7. The method for UAV cruise path planning with multi-priority dynamic coverage according to claim 1 is characterized in that: The scheme evaluation of all cruise path planning schemes includes redundancy evaluation and priority evaluation. The calculation formula of the redundancy evaluation is: ； in, Plan a cruise route The redundancy score of For the cruise path The path length, is the total number of cruise grids; The calculation formula for the priority assessment is: ； in, Plan a cruise route The priority score of For the cruise grid The traversal cycle, the lower the score, the higher the priority.

8. The method for UAV cruise path planning with multi-priority dynamic coverage according to claim 7, characterized in that: The cruise path planning scheme with the lowest redundancy score is selected as the optimal cruise path planning scheme. If there are multiple cruise path planning schemes with the lowest redundancy scores, the cruise path planning scheme with the lowest priority score is selected as the optimal cruise path planning scheme. The number of cruise paths in the optimal cruise path planning scheme is the number of drones required for the cruise.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images