Dynamic arrival-oriented multi-unmanned aerial vehicle and vehicle collaborative data acquisition method and system

Abstract

The invention relates to a multi-unmanned aerial vehicle and vehicle collaborative data acquisition method and system for dynamic arrival, and the method comprises the steps: firstly constructing an acquisition model comprising a vehicle, unmanned aerial vehicles and a data center, then constructing a global data acquisition and adjustment scheme, finally executing the global data acquisition scheme, and executing the adjustment scheme when a new task arrives; the global data acquisition scheme comprises parking node selection, data acquisition node distribution, vehicle route and flight route planning and distribution and other schemes, and the adjustment scheme comprises newly added data acquisition node division, local and global data acquisition and other schemes; through the advantages of flexibility of the unmanned aerial vehicle, high endurance of the vehicle and the like, the target area is decomposed, meanwhile, multiple acquisition schemes are constructed to cover various different conditions in the data acquisition process, newly-added tasks are adjusted in real time through the adjustment schemes, and the task dynamic arrival speed is improved under the scene that the tasks arrive dynamically. The multiple unmanned aerial vehicles and the vehicles cooperatively and efficiently collect data, and the adaptive capacity of the collection scheme is improved.

Description

Technical Field

[0001] The present invention relates to data collection means, belonging to the technical field of data collection, and in particular to a method and system for collaborative data collection of multiple unmanned aerial vehicles and vehicles facing dynamic arrival. Background Art

[0002] In recent years, with the rapid development of science and technology, drone technology has been significantly improved and has gradually been maturely applied in many fields such as economy, agriculture, and national defense, driving these fields to develop in a more intelligent direction. For example, in agriculture, drones are used to collect data on crop growth; traditional manual collection methods are not only inefficient, but also difficult to reach certain areas due to geographical restrictions, which poses certain risks. In contrast, drones equipped with cameras can overcome geographical obstacles and efficiently capture the growth of crops at designated collection points, improving the efficiency and safety of data collection.

[0003] Compared with a single drone, multiple drones show significant advantages and greater potential when performing data collection tasks. First, multiple drones can share tasks and improve data collection efficiency. For example, in the task of collecting crop growth data, multiple drones can cover different areas at the same time, thereby completing the collection of large-scale crop growth data in a shorter time. Secondly, multiple drones are more reliable and robust, which means that even if individual drones fail, other drones can continue to perform tasks to ensure the continuity and integrity of data collection; in addition, multiple drones are highly flexible and adaptable, and can dynamically adjust the number and flight paths of drones according to mission requirements to cope with complex environments and emergencies. However, due to the limited endurance of drones, there are certain limitations when performing large-area tasks. In order to overcome this problem, vehicles are usually used as mobile base stations to carry multiple drones to perform tasks together, thereby improving the coverage and execution effect of the task; however, the acquisition schemes of the prior art mainly focus on the scheduling of drones in static task scenarios, that is, the task scenarios are fixed, which leads to poor adaptability of the scheduling schemes, thereby affecting the collection efficiency. However, in actual collection tasks, the task scenarios often have temporary changes due to traffic control and other situations, that is, in the scenario of dynamic arrival of the task, the target area may randomly add data collection nodes during the execution of the task, resulting in the need for subsequent data collection schemes to consider access to the newly added data collection nodes. Therefore, there is an urgent need for a data collection method with better adaptability and higher efficiency to solve the above-mentioned defects in the prior art. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects and problems of poor adaptability and low efficiency in the prior art, and to provide a multi-UAV and vehicle collaborative data collection method and system for dynamic arrival with good adaptability and high efficiency.

[0005] To achieve the above objectives, the technical solution of the present invention is: a method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival, comprising:

[0006] S1. Constructing a data collection model for multi-UAV and vehicle collaboration, wherein the data collection model includes at least one vehicle, several UAVs and a data center;

[0007] S2. Divide the target area into several node areas and set parking nodes and data collection nodes in each node area; with the goal of minimizing the data collection completion time, construct an initial global data collection plan and data collection adjustment plan;

[0008] S3, perform data collection tasks; the data collection tasks are as follows:

[0009] The vehicle starts from the data center and drives to the first parking node. The drone takes off from the vehicle and drives to the data collection node to collect data. The drone returns to the vehicle and the vehicle drives to the next parking node. This cycle repeats until the vehicle traverses all node areas and returns to the data center, completing the data collection task.

[0010] The initial global data collection scheme includes: parking node selection and data collection node allocation scheme, vehicle route solution scheme, drone flight route solution scheme at parking nodes, and drone flight route allocation scheme;

[0011] The steps for constructing the initial global data acquisition scheme are as follows:

[0012] S21, based on the parking node selection and data collection node allocation scheme, by calculating the average distance between each candidate parking node and the data collection nodes it covers, the candidate parking nodes with the smallest average distance are selected in turn as parking nodes, and each data collection node is allocated to the nearest parking node, so as to determine the parking nodes and allocate the data collection nodes;

[0013] S22, for the parking nodes determined in step S21, solving the driving route of the vehicle, the vehicle route solving scheme is based on the ant colony algorithm, randomly determining the starting parking node of the ants, continuously obtaining the ant positions and the unvisited parking nodes, calculating the access probability to determine the next parking node, selecting and recording the shortest vehicle route, so as to determine the shortest vehicle driving route of the vehicle;

[0014] S23, for the data collection nodes assigned to each parking node determined in step S21, the flight route of the drone at the parking node is solved by the solution, and the steps are as follows:

[0015] S231, memory pool initialization: putting the flight route of the drone into the memory pool, deleting the data collection nodes included in the flight route of the drone from the data collection node set, and initializing the data collection node set until the maximum capacity of the memory pool is reached;

[0016] S232, UAV flight route solution construction: Use the average access distance to evaluate the UAV flight routes in the memory pool, and select the flight route with the shortest average access distance to construct a solution;

[0017] S233, optimization of drone flight route solution: for the drone flight route solution constructed in step S232, two drone flight routes are randomly selected from the solution, and two data collection nodes with the closest distance are found to exchange and migrate nodes to obtain several new route selection pairs, and the shortest drone flight route is selected from the new route selection pairs to replace the two randomly selected drone flight routes as the optimized drone flight route solution;

[0018] S234, memory pool update: put the optimized UAV flight route in the UAV flight route solution into the memory pool, and delete the poor UAV flight route in the memory pool;

[0019] S24, for the optimized UAV flight route solution in step S233, the UAV flight route allocation scheme sorts all UAV flight routes and allocates the longest unallocated UAV flight route to the UAV with the shortest total length of the currently allocated route;

[0020] When a new task arrives during the execution of the above data collection task, the data collection adjustment plan is executed to adjust the current data collection plan;

[0021] The data collection adjustment scheme includes: a new data collection node division scheme, a local data collection solution scheme, and a global data collection scheme; the steps for constructing the data collection adjustment scheme are as follows:

[0022] S25, the newly added data collection node division scheme defines an unvisited area and an unplanned area, divides the newly added data collection nodes distributed in the unvisited area into an unvisited data collection node set, and divides the newly added data collection nodes distributed in the unplanned area into an unplanned data collection node set;

[0023] S26: For the unvisited data collection node set and the unplanned data collection node set in step S25, there are two situations:

[0024] The first type: the drone is collecting data and the vehicle is parked at the parking node; the local data collection solution is solved by solving the drone flight path solution at the parking node, solving the drone flight paths of the current parking node and the remaining unvisited parking nodes, and assigning the flight paths to each drone;

[0025] The second type: all drones are on the vehicle, and the vehicle is driving; then the local data collection solution solves the flight routes of the drones at the remaining unvisited parking nodes through the flight route solution of the drones at the parking nodes, and allocates the flight routes to each drone;

[0026] S27, during the execution of the local data collection solution in step S26, when the number of data collection nodes in the unplanned data collection node set is greater than the number of data collection nodes in the unvisited data collection node set, the global data collection solution is executed, and the steps are as follows:

[0027] S271, merging the data collection nodes in the unvisited data collection node set and the data collection nodes in the unplanned data collection node set to obtain a global data collection node set;

[0028] S272. Solve the global data collection node set through parking node selection and data collection node allocation scheme to obtain parking nodes and data collection nodes allocated to each parking node; solve the shortest vehicle driving route through the greedy algorithm; for each parking node, obtain the UAV flight route at the parking node through the UAV flight route solution at the parking node; for each parking node, allocate a flight route for the UAV through the UAV flight route allocation scheme.

[0029] The steps of the parking node selection and data collection node allocation scheme are as follows:

[0030] A1, taking the data collection node set C and the candidate parking node set A as input, and taking the parking node set P and the data collection node set C assigned to each parking node P is the output;

[0031] A2. Calculate the average distance between each candidate parking node and the data collection nodes covered within the maximum flight distance, and sort the parking nodes in ascending order according to the average distance to construct a set of data collection nodes covered by the candidate parking nodes;

[0032] The expression of the average distance is as follows:

[0033]

[0034] in: is the candidate parking node a iWith data collection node c j The distance between

[0035] Then for candidate parking node a i The set of data collection nodes covered It is expressed as follows:

[0036]

[0037] Where: d uav is the maximum flight distance of the drone;

[0038] A3. Select the candidate parking node with the smallest average distance as the parking node, remove the candidate parking node from the candidate parking node set A; remove the data collection node covered by the candidate parking node from the data collection node set C; add the candidate parking node to the parking node set P; repeat the above operation until the data collection node set C is empty;

[0039] A4. For each data collection node, find the parking node closest to it and assign the data collection node to the parking node; the expression is as follows:

[0040] C pi =C pi ∪{c i};

[0041] Where: C pi The data collection node assigned to the parking node pi, c i is the i-th data collection node in C.

[0042] The steps of the vehicle routing solution are as follows:

[0043] B1, based on the vehicle route node set r truck =P∪ n0 As input, the shortest vehicle route r best is the output; where: n0 is the data center node;

[0044] B2. Randomly generate the starting parking node for each ant and record it in the vehicle route table path_table;

[0045] B3. Get the parking node n where the ant is currently located visting And the unvisited parking node list unvisited_list;

[0046] B4. Calculate the access probability of each unvisited parking node and determine the next parking node n to be visited. next ;

[0047] B5. Update the vehicle route table path_table, the unvisited parking node list unvisited_list and the current ant location n visting ←n next ;

[0048] If there are still parking nodes that the current ant has not visited, return to step B4; if there are still untraversed ants, return to step B3; if all ants have been traversed and there are no unvisited parking nodes for the current ant, proceed to step B6;

[0049] B6. Get the shortest vehicle route r in the vehicle route table path_table shortest , and record the shortest vehicle route R in this iteration shortest =R shortest ∪{r shortest}, then calculate the pheromone increment on the vehicle route and update the pheromone matrix M phe；

[0050] B7. Iterate according to steps B2-B6 until the maximum number of iterations is reached, and the shortest vehicle route r shortest Get the shortest vehicle route r best .

[0051] The steps of the UAV flight path solution construction are as follows:

[0052] C1, takes the memory pool M, parking node pi and the data collection node set Cpi allocated by parking node pi as input, and takes the flight route Rpi of the drone at parking node pi as output;

[0053] C2. Sort the UAV flight routes in the memory pool in ascending order according to the average access distance of the UAV flight routes; the expression of the average access distance is as follows:

[0054]

[0055] Where; l(r) is the length of the UAV flight route, |r| is the number of data collection nodes in the UAV flight route;

[0056] C3. Select a UAV flight route with the smallest average access distance and add it to R pi and removed from memory pool M;

[0057] C4, the data collection nodes in the current UAV's flight path are allocated from the parking node pi to the data collection node set C pi Removed;

[0058] C5. Remove the data collection nodes in the remaining UAV flight routes in the memory pool M that are the same as the current UAV flight route; if C pi If it is not empty, return to step C2 until C pi Is empty.

[0059] The steps of optimizing the UAV flight path are as follows:

[0060] D1, taking the parking node pi and the UAV flight path solution Rpi at the parking node pi as input, and the optimized UAV flight path solution Rpi pi,new is the output;

[0061] D2. Solve the flight path of the drone from the parking node pi. pi Randomly select two UAV flight routes r1 and r2, and find the two data collection nodes C closest to r1 and r2. i and c j ;

[0062] D3, the UAV flight routes r1, r2 and data collection node c i 、c j Perform data collection node exchange and data collection node migration to obtain several new route pairs, and put the several new route pairs and drone flight routes r1 and r2 into the route group;

[0063] D4. Remove the route pairs whose route length exceeds the maximum flight distance of the drone in the route group, and take out the route pair with the longest flight distance and the shortest flight time in the route group, and use this route pair to replace R pi The optimized UAV flight path solution R is obtained by using r1 and r2 in pi，new ; Return to step D2 until the maximum number of iterations is reached.

[0064] The steps of the UAV flight route allocation scheme are as follows:

[0065] F1. Solve R with optimized drone flight path pi，new And the set of drones U that are not assigned with drone flight routes is taken as input to the set of drones U that are assigned with drone flight routes new is the output;

[0066] E2. According to the length l(r) of the UAV flight path, R pi，new Sort the UAV flight routes in descending order;

[0067] E3. Take R pi,new The longest drone flight route in l(r) j , get the time t required for the drone in U to complete the currently assigned task uThe smallest drone k , then r j Assign to u k , and updates the total time the drone spends accessing the data collection node at the parking node pi At the same time, r j From R pi,new Removed;

[0068] If at this time R pi，new If it is not empty, return to step E3 until R pi，new If it is empty, get the set of drones U that assign the flight routes of drones new .

[0069] The steps of the newly added data collection node division scheme are as follows:

[0070] If a new data collection node is added during the UAV's data collection mission, the data collection node cj needs to be divided at this time. The steps are as follows:

[0071] F1. Define the area covered by the remaining unvisited parking nodes as the unvisited area Ω uv , the remaining area is the unplanned areaΩ up ;

[0072] F2. Define the unvisited area Ω uv The data collection nodes inside are the unvisited data collection node set C uv , unplanned areaΩ up The data collection nodes inside are the unplanned data collection node set C up ;

[0073] F3. Distribute the newly added data collection nodes in Ω uv With Ω up For the distribution in Ω uv The data collection nodes in C uv For the distribution in Ω up The data collection nodes in C up middle;

[0074] F4, for unvisited areasΩ uv ,UAV flight routes are allocated using a local data collection solution;

[0075] The steps of the local data collection solution are as follows:

[0076] G1, parking node set P, data collection node C assigned to each parking node P , the set of drones U that are not assigned with drone flight routes is taken as input, and the set of drones U that are assigned with drone flight routes is taken as input. new is the output;

[0077] G2. If the vehicle is parked at a parking node, obtain the parking node p where the vehicle is parked. now , and p now Delete from P; adjust the current route of the drone based on the current route adjustment plan of the drone, and p based on the solution of the flight route of the drone at the parking node now The UAV flight route at the location is solved; based on the UAV flight route allocation scheme, p now The UAV flight routes at the locations are assigned to the UAVs;

[0078] If the vehicle is traveling at this time, the UAV flight route is solved for each parking node based on the UAV flight route solution at the parking node, and the UAV flight route is allocated to the UAV based on the UAV flight route allocation solution for each parking node.

[0079] The steps of the current route adjustment plan of the drone are as follows:

[0080] H1, parking node pi, the remaining unvisited data collection node set E at parking node pi pi , the positions and remaining range of all drones are used as input; the adjusted current route of each drone is used as output;

[0081] H2, from E pi Find the nearest drone u from the positions of all drones j and data collection node cj;

[0082] H3. If the drone u j After visiting the data collection node c j After that, the vehicle can be returned smoothly, then c j Join UAV j In the current route, c j From E pi Remove and update the drone j Otherwise, the drone will be j The battery life is set to 0;

[0083] H4. If the endurance of all drones is not 0 and E pi If it is not empty, return to step H2 until all drones or E are updated. pi Is empty.

[0084] The steps of the global data acquisition scheme are as follows:

[0085] K1, collect node set C with unvisited data uv , Unplanned data collection node set C up, the candidate parking node set A is input, and the shortest vehicle driving route r best , assigning drone flight routes to a drone set U new is the output;

[0086] K2, collect the unvisited data node set C uv and the unplanned data collection node set C up The data collection nodes in are merged to obtain the global data collection node set C;

[0087] K3, call the parking node selection and data collection node allocation algorithm to obtain the parking node pi and the data collection node set C assigned to each parking node P ;

[0088] K4, based on the greedy algorithm to obtain the shortest vehicle route r best ;

[0089] K5. For each parking node, based on the solution for the UAV flight route at the parking node, the UAV flight route at the parking node is solved; for each parking node, based on the UAV flight route allocation plan, the UAV set to which the UAV flight route is allocated is obtained.

[0090] A multi-UAV and vehicle collaborative data acquisition system for dynamic arrival, the system is applied to the above method, the system comprising:

[0091] A data collection model building module, used to build a data collection model for multi-UAV and vehicle collaboration, wherein the data collection model includes at least one vehicle, several UAVs and a data center;

[0092] The data collection and adjustment plan building module is used to divide the target area into several node areas and set parking nodes and data collection nodes in each node area; with the goal of minimizing the data collection completion time, the initial global data collection plan and data collection adjustment plan are built;

[0093] The initial global data collection scheme includes: parking node selection and data collection node allocation scheme, vehicle route solution scheme, drone flight route solution scheme at parking nodes, and drone flight route allocation scheme;

[0094] The steps for constructing the initial global data acquisition scheme are as follows:

[0095] S21, based on the parking node selection and data collection node allocation scheme, by calculating the average distance between each candidate parking node and the data collection nodes it covers, the candidate parking nodes with the smallest average distance are selected in turn as parking nodes, and each data collection node is allocated to the nearest parking node, so as to determine the parking nodes and allocate the data collection nodes;

[0096] S22, for the parking nodes determined in step S21, solving the driving route of the vehicle, the vehicle route solving scheme is based on the ant colony algorithm, randomly determining the starting parking node of the ants, continuously obtaining the ant positions and the unvisited parking nodes, calculating the access probability to determine the next parking node, selecting and recording the shortest vehicle route, so as to determine the shortest vehicle driving route of the vehicle;

[0097] S23, for the data collection nodes assigned to each parking node determined in step S21, the flight route of the drone at the parking node is solved by the solution, and the steps are as follows:

[0098] S231, memory pool initialization: putting the flight route of the drone into the memory pool, deleting the data collection nodes included in the flight route of the drone from the data collection node set, and initializing the data collection node set until the maximum capacity of the memory pool is reached;

[0099] S232, UAV flight route solution construction: Use the average access distance to evaluate the UAV flight routes in the memory pool, and select the flight route with the shortest average access distance to construct a solution;

[0100] S233, optimization of drone flight route solution: for the drone flight route solution constructed in step S232, two drone flight routes are randomly selected from the solution, and two data collection nodes with the closest distance are found to exchange and migrate nodes to obtain several new route selection pairs, and the shortest drone flight route is selected from the new route selection pairs to replace the two randomly selected drone flight routes as the optimized drone flight route solution;

[0101] S234, memory pool update: put the optimized UAV flight route in the UAV flight route solution into the memory pool, and delete the poor UAV flight route in the memory pool;

[0102] S24, for the optimized UAV flight route solution in step S233, the UAV flight route allocation scheme sorts all UAV flight routes and allocates the longest unallocated UAV flight route to the UAV with the shortest total length of the currently allocated route;

[0103] The data collection task execution module is used to execute the data collection task and, when a new task arrives during the data collection task execution, execute the data collection adjustment plan to adjust the current data collection plan;

[0104] The data collection tasks are as follows:

[0105] The vehicle starts from the data center and drives to the first parking node. The drone takes off from the vehicle and drives to the data collection node to collect data. The drone returns to the vehicle and the vehicle drives to the next parking node. This cycle repeats until the vehicle traverses all node areas and returns to the data center, completing the data collection task.

[0106] The data collection adjustment scheme includes: a new data collection node division scheme, a local data collection solution scheme, and a global data collection scheme;

[0107] The steps for constructing the data collection and adjustment scheme are as follows:

[0108] S25, the newly added data collection node division scheme divides the newly added data collection nodes distributed in the unvisited area into an unvisited data collection node set and divides the newly added data collection nodes distributed in the unplanned area into an unplanned data collection node set by defining an unvisited area and an unplanned area;

[0109] S26: For the unvisited data collection node set and the unplanned data collection node set in step S25, there are two situations:

[0110] The first type: the drone is collecting data and the vehicle is parked at the parking node; the local data collection solution is solved by solving the drone flight path solution at the parking node, solving the drone flight paths of the current parking node and the remaining unvisited parking nodes, and assigning the flight paths to each drone;

[0111] The second type: all drones are on the vehicle, and the vehicle is driving; then the local data collection solution solves the flight routes of the drones at the remaining unvisited parking nodes through the flight route solution of the drones at the parking nodes, and allocates the flight routes to each drone;

[0112] S27, during the execution of the local data collection solution in step S26, when the number of data collection nodes in the unplanned data collection node set is greater than the number of data collection nodes in the unvisited data collection node set, the global data collection solution is executed, and the steps are as follows:

[0113] S271, merging the data collection nodes in the unvisited data collection node set and the data collection nodes in the unplanned data collection node set to obtain a global data collection node set;

[0114] S272. Solve the global data collection node set through parking node selection and data collection node allocation scheme to obtain parking nodes and data collection nodes allocated to each parking node; solve the shortest vehicle driving route through the greedy algorithm; for each parking node, obtain the UAV flight route at the parking node through the UAV flight route solution at the parking node; for each parking node, allocate a flight route for the UAV through the UAV flight route allocation scheme.

[0115] Compared with the prior art, the present invention has the following beneficial effects:

[0116] The present invention discloses a method and system for collaborative data collection of multiple unmanned aerial vehicles and vehicles facing dynamic arrival. The method first constructs a collection model including vehicles, unmanned aerial vehicles and data centers, then constructs a global data collection plan and an adjustment plan, and finally executes the global data collection plan to perform collection tasks, and executes the adjustment plan when a new task arrives; the global data collection plan includes parking node selection and data collection node allocation, vehicle routes, unmanned aerial vehicle flight routes, flight route allocation and other plans, and the adjustment plan includes new data collection node division, local data collection, global data collection and other plans; in application, the present invention uses the advantages of the flexibility of unmanned aerial vehicles and the high endurance of vehicles, and decomposes a large target area into small target areas centered on parking nodes, and constructs a variety of collection plans to cover different situations such as parking, data collection, route allocation and other situations in the data collection process, and adjusts the new tasks in real time through the adjustment plan, thereby realizing efficient data collection of multiple unmanned aerial vehicles and vehicles in a scenario of dynamic arrival of tasks, and improving the adaptability of the collection plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0117] Figure 1 It is a flow chart of the method steps of the present invention.

[0118] Figure 2 It is a structural schematic diagram of the data acquisition model of the present invention.

[0119] Figure 3 It is a schematic diagram of the system structure of the present invention.

[0120] Figure 4 It is a schematic diagram of the device structure of the present invention.

[0121] In the figure: data acquisition model construction module 1, data acquisition and adjustment plan construction module 2, data acquisition task execution module 3, processor 4, memory 5, computer program code 51. DETAILED DESCRIPTION

[0122] The present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0123] Embodiment 1:

[0124] See also Figure 1 , a method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival, comprising:

[0125] S1. Constructing a data collection model for multi-UAV and vehicle collaboration, wherein the data collection model includes at least one vehicle, several UAVs and a data center;

[0126] S2. Divide the target area into several node areas and set parking nodes and data collection nodes in each node area; with the goal of minimizing the data collection completion time, construct an initial global data collection plan and data collection adjustment plan;

[0127] S3, executing the data collection task, and when a new task arrives during the data collection task execution, executing the data collection adjustment plan to adjust the current data collection plan;

[0128] The data collection tasks are as follows:

[0129] The vehicle starts from the data center and drives to the first parking node. The drone takes off from the vehicle and drives to the data collection node to collect data. The drone returns to the vehicle and the vehicle drives to the next parking node. This cycle repeats until the vehicle traverses all node areas and returns to the data center, completing the data collection task.

[0130] In the application, this solution designs a model for collaborative data collection between multiple drones and vehicles, using the idea of ​​dividing a large area into multiple small areas for regional collection. Specifically, the entire target area is divided into small areas according to the selected parking nodes, and the data collection work is divided into two steps: the vehicle driving to the parking node and the drone taking off from the vehicle to collect data. In this mode, only the problem of vehicle and drone path planning and drone flight route allocation needs to be solved, and drones and vehicles can collaborate and autonomously complete the data collection task.

[0131] Further, such as Figure 2As shown, in this solution, the vehicle is the base station for the drone to move. The vehicle's hardware and software settings have the following features: it can carry multiple drones and drone batteries so that the batteries can be replaced immediately after the drone completes the task; it has the ability to receive and send data, can receive new tasks from the data center, and can transmit the data collected by the drone to the data center in real time; it has independent computing and control capabilities, can calculate the paths of the vehicle and drone according to the tasks and the tasks assigned by the drone, and can control the drone to collect data along the specified path. The drone itself has the function of taking pictures and storing data, and works under the command of the drone central control system on the vehicle; the drone can perform data collection tasks according to the pre-planned flight route; when the battery life is insufficient, it can fly back to the vehicle to replace the battery.

[0132] The vehicle's driving route can be described as an ordered sequence of departure nodes, parking nodes, and return nodes; when the vehicle departs from the data center, the departure node and return node of the vehicle's driving route are both data center n0; when new data collection nodes are added during task execution and global replanning is required, the departure node of the vehicle's driving route is the vehicle's current location, and the return node is still data center node n0. Therefore, the node set of the vehicle's driving route can be recorded as r truck ={n truck , p i ,…,p j , n0}, where n truck is the current location of the vehicle, which can be a data center or a node in the target area; i is the i-th parking node; p j is the jth parking node; define the decision variable x i，j = {0, 1}, when a drone or vehicle moves from node i to node j in a route, x i，j =1; otherwise, x i，j = 0. Based on this decision variable, the total length of the vehicle's route l(r trucck ) is expressed as follows:

[0133]

[0134] Where: d i，j is the straight-line distance between parking node i and parking node j;

[0135] The time taken by the vehicle to travel is expressed as follows:

[0136]

[0137] Where: v truck is the vehicle speed;

[0138] When a vehicle stops at a parking node, multiple drones will be released to access the data collection nodes assigned to the parking node. At this time, it is necessary to reasonably divide these data collection nodes into multiple drone flight routes and reasonably and evenly distribute all drone flight routes to each drone. In this way, when multiple drones perform tasks in parallel, all data collection nodes at the parking node can be accessed in the shortest time.

[0139] A UAV flight route can be described as an ordered sequence of departure nodes, data collection nodes, and return nodes; when the UAV flies out from the vehicle parking node to collect data, the departure node and return node of the UAV flight route are both the vehicle parking node. When the UAV is collecting data and the task arrives dynamically, once a new data collection node is added to the area where the parking node is located, the current flight route of the UAV needs to be adjusted according to the real-time location of the UAV and the remaining battery life. At this time, the departure node of the UAV is the current location of the UAV, and the return node is still the parking node.

[0140] In this scheme, a series of decision variables are defined. Decision variables are unknown quantities in the model. By selecting appropriate values ​​of decision variables, the objective function can obtain the optimal value while satisfying a series of constraints.

[0141] Define the decision variable y i，j ={0,1}, when data collection node c j is divided into parking nodes p i When i，j =1; otherwise, y i，j =0; then it is divided into parking node p i The data collection nodes at are represented as follows:

[0142] C pi ={c j |c j ∈C,y i，j =1};

[0143] Define the decision variable S j，k = {0, 1}, when the parking node p i The data collection node G k Assigned to the jth drone flight route When S j，k =1; otherwise, S j，k = 0. In addition to the data collection nodes, the node set in the UAV flight route also includes the departure node and return node of the UAV; after completing a route, the UAV needs to go to the vehicle to recharge, so the return node of the UAV can only be the parking node p i ; The starting node of the drone can be the parking node pi , or the current position of the drone n uav ; then the parking node p i The jth UAV flight path at The node set in is represented as follows:

[0144]

[0145] The length of a flight route of a drone is expressed as follows:

[0146]

[0147] Where: d k，m is the straight-line distance between data acquisition node k and data acquisition node m; x k，m When it is 1, it means it is on the UAV flight route. In the data collection node m, the data collection node k is behind, x k，m When it is 0, it means it is on the UAV flight route. In the data collection node m, it is not behind the data collection node k;

[0148] Note the parking node p i The set of UAV flight routes is R pi , define the decision variable z k，j = {0, 1}, when the parking node p i When the jth UAV flight route is assigned to UAV k, z k，j =1; otherwise, z k，j = 0; then the set of UAV flight routes assigned to UAV k at parking node pi is expressed as follows:

[0149]

[0150] Parking node p i The total flight distance l of UAV k uk It is expressed as follows:

[0151]

[0152] Then drone k is at parking node p i The total time t spent on accessing the data collection node uk It is expressed as follows:

[0153]

[0154] The vehicle parking time at parking node pi is expressed as follows:

[0155]

[0156] Finally, the goal of this scheme is to minimize the completion time of a global data collection scheme; its objective function is expressed as follows:

[0157]

[0158] During the execution of the data collection task, based on the initial global data collection plan, the data collection plan is optimized with the goal of minimizing the data collection completion time. When a new task arrives during the data collection task, the data collection plan adjustment plan is executed to adjust the current data collection plan.

[0159] The initial global data collection scheme includes: parking node selection and data collection node allocation scheme, vehicle route solution scheme, drone flight route solution scheme at parking nodes, and drone flight route allocation scheme;

[0160] The steps for constructing the initial global data acquisition scheme are as follows:

[0161] S21, based on the parking node selection and data collection node allocation scheme, by calculating the average distance between each candidate parking node and the data collection nodes it covers, the candidate parking nodes with the smallest average distance are selected in turn as parking nodes, and each data collection node is allocated to the nearest parking node, so as to determine the parking nodes and allocate the data collection nodes;

[0162] This scheme is aimed at the parking node selection and data collection node allocation scheme. The scheme aims to select a part of the candidate parking nodes as the final parking nodes. It follows two principles: covering all data collection nodes and minimizing the distance between the parking nodes and the data collection nodes they cover to ensure the comprehensiveness and efficiency of data collection. The scheme adopts a greedy strategy. First, the average distance between each candidate parking node and the data collection nodes it covers is calculated, and then the candidate parking node with the smallest average distance is selected as the parking node; the selected candidate parking node and the data collection nodes it covers are deleted from their respective node sets, and the average distance of the remaining candidate parking nodes is recalculated, and the above process is repeated until all data collection nodes are covered. Finally, each data collection node is assigned to the nearest parking node; through this method of gradually selecting the optimal parking node, it is ensured that the average distance can be effectively reduced at each step to achieve efficient data collection coverage.

[0163] Furthermore, the steps of the parking node selection and data collection node allocation scheme are as follows:

[0164] A1, taking the data collection node set C and the candidate parking node set A as input, and taking the parking node set P and the data collection node set C assigned to each parking node POutput: Initialize the parking node set P = {} and the data collection node set C assigned to each parking node P ={C p1 ={},C p2 ={},...,C p|CP |={}};

[0165] A2. Calculate the average distance between each candidate parking node and the data collection nodes covered within the maximum flight distance, and sort the parking nodes in ascending order according to the average distance to construct a set of data collection nodes covered by the candidate parking nodes;

[0166] The expression of the average distance is as follows:

[0167]

[0168] in: is the candidate parking node a i With data acquisition node c j The distance between

[0169] Then for candidate parking node a i The covered data collection node set C ai It is expressed as follows:

[0170]

[0171] Where: d uav is the maximum flight distance of the drone;

[0172] A3. Select the candidate parking node with the smallest average distance as the parking node, remove the candidate parking node from the candidate parking node set A; remove the data collection node covered by the candidate parking node from the data collection node set C; add the candidate parking node to the parking node set P; the expressions of the above three steps are as follows: A = A / a i , P=P∪{a i}; If the data collection node set C is not empty, return to step A2; repeat the above operation until the data collection node set C is empty;

[0173] A4. For each data collection node, find the parking node closest to it and assign the data collection node to the parking node; the expression is as follows:

[0174] C pi =C pi ∪{c i};

[0175] Where: C pi The data collection node assigned to the parking node pi, ci is the i-th data collection node in C.

[0176] S22, for the parking nodes determined in step S21, solving the driving route of the vehicle, the vehicle route solving scheme is based on the ant colony algorithm, randomly determining the starting parking node of the ants, continuously obtaining the ant positions and the unvisited parking nodes, calculating the access probability to determine the next parking node, selecting and recording the shortest vehicle route, so as to determine the shortest vehicle driving route of the vehicle;

[0177] In this solution, the parking nodes have been solved in step S21, so this step needs to solve the vehicle's driving route based on the parking nodes. For solving the vehicle route, the traveling salesman problem (TSP) model in graph theory is usually used to solve it. The traveling salesman problem describes a situation where a traveling salesman starts from a certain city, passes through all cities once, and finally returns to the starting city, and requires the total length of the entire travel route to be the shortest; in the initial global data collection solution, the vehicle starts from the data center, passes through all the parking nodes, and finally returns to the data center. Its optimization goal is to make the total length of the vehicle's driving route the shortest. In order to solve this problem, the present invention adopts the Ant Colony Algorithm (ACA).

[0178] Furthermore, the steps of the vehicle route solution are as follows:

[0179] B1, based on the vehicle route node set r truck =P∪ n0 As input, the shortest vehicle route r best is the output; where: n0 is the data center node; initialize the parking node distance matrix M d , number of parking nodes n, number of ants m, pheromone volatility coefficient, pheromone importance, heuristic function importance, total pheromone release, maximum number of iterations epoch, shortest vehicle route R per iteration shortest ={}, pheromone matrix M phe And other parameters;

[0180] B2. Randomly generate the starting parking node for each ant and record it in the vehicle route table path_table;

[0181] B3. Get the parking node n where the ant is currently located visting And the unvisited parking node list unvisited_list;

[0182] B4. Calculate the access probability of each unvisited parking node and determine the next parking node n to be visited. next ;

[0183] B5. Update the vehicle route table path_table, the unvisited parking node list unvisited_list and the current ant location n visting ←n next ;

[0184] If there are still parking nodes that the current ant has not visited, return to step B4; if there are still untraversed ants, return to step B3; if all ants have been traversed and there are no unvisited parking nodes for the current ant, proceed to step B6;

[0185] B6. Get the shortest vehicle route r in the vehicle route table path_table shortest , and record the shortest vehicle route R in this iteration shortest =R shortest ∪{r shortest}, then calculate the pheromone increment on the vehicle route and update the pheromone matrix M phe ;

[0186] B7. Iterate according to steps B2-B6 until the maximum number of iterations is reached, and the shortest vehicle route r shortest Get the shortest vehicle route r besst .

[0187] S23, for each parking node that has been determined in step S21 to be assigned a data collection node; this step needs to solve the UAV flight path according to each parking node and the data collection node assigned to the parking node. For solving the UAV flight path at the parking node, this solution adopts the AMP algorithm. The UAV flight path solution solution at the parking node solves the UAV flight path, and its steps are as follows:

[0188] S231, memory pool initialization: putting the flight route of the drone into the memory pool, deleting the data collection nodes included in the flight route of the drone from the data collection node set, and initializing the data collection node set until the maximum capacity of the memory pool is reached;

[0189] In this solution, the memory pool is initialized to store the flight routes of the drone. The initialization process uses a random + greedy algorithm to construct multiple drone flight routes. Specifically, for a parking node pi, a data collection node is randomly selected from the data collection node set Cpi assigned to pi as the first data collection node to be visited in a drone flight route, and then the data collection node closest to the current data collection node is selected as the next node, and this is repeated until the length of the constructed drone flight route exceeds the maximum flight distance of the drone. The constructed drone flight route is put into the memory pool, and the data collection nodes contained in the drone flight route are removed from Cpi. piRepeat this process until C pi All data collection nodes are divided into UAV flight routes. Then initialize C pi , continue to build until the maximum capacity n of the memory pool is reached max .

[0190] Furthermore, the steps of the memory pool initialization scheme are as follows:

[0191] N1, data collection node set C allocated by parking nodes pi, pi pi as input and memory pool M as output; initialize memory pool M = {} and the maximum capacity of memory pool n max ;

[0192] L2, the data collection node set C assigned to pi pi Save as temporary data collection node set C pi _temp;

[0193] L3, initialize a drone flight route to be empty, and the remaining endurance of the drone is d uav ; Its expression is: r = {}, dist = d uav ; Among them: r represents the currently constructed UAV flight route, dist represents the control parameter for judging whether the current route can be added with data collection nodes, d uav Represents the maximum flight distance of the drone;

[0194] L4, select C pi A random data collection node C in _temp j , add the data collection node to r and select pi _temp, and then record the current position of the drone c now And update dist; its expression is: dist = dist-d pi，cj ; Among them: d pi，cj Indicates parking node pi to data collection node c j The distance between

[0195] L5, select the distance C from the current position of the drone now The nearest data collection node C j , and update dist; its expression is: in: is the current position of the drone c now To data collection node c j The distance between

[0196] At this time, if dist>0, then add the data collection node to r and pi _temp is removed; if Cpi_temp is not empty and dist>0, then return to step L5; if the number of flight routes of drones in memory pool M is less than n max , then return to step L2; repeat the above steps until C pi _temp is empty.

[0197] S232, UAV flight route solution construction: Use the average access distance to evaluate the UAV flight routes in the memory pool, and select the flight route with the shortest average access distance to construct a solution;

[0198] In this solution, for the construction of the UAV flight route solution, it is necessary to select the UAV flight route from the initialized memory pool to construct the solution of the UAV flight route at the current parking node. First, the UAV flight routes in the memory pool are evaluated, and the average access distance is used to evaluate each UAV flight route in the memory pool. Then, the route with the shortest average access distance is selected to construct the solution, and then the remaining UAV flight routes in the memory pool are updated. If the remaining routes contain data collection nodes in the selected routes, these data collection nodes are removed and the average access distance of the remaining routes is recalculated until the selected route contains all the data collection nodes in the area.

[0199] Furthermore, the steps of deconstructing the UAV flight path are as follows:

[0200] C1, memory pool M, parking node pi and data collection node set C allocated by parking node pi pi As input, the flight path Rpi of the UAV at the parking node pi is output;

[0201] C2. Sort the UAV flight routes in the memory pool in ascending order according to the average access distance of the UAV flight routes; the expression of the average access distance is as follows:

[0202]

[0203] Where; l(r) is the length of the UAV flight route, |r| is the number of data collection nodes in the UAV flight route;

[0204] C3, select a drone flight route with the shortest average access distance, add it to Rpi, and remove it from the memory pool M;

[0205] C4, the data collection nodes in the current UAV's flight route r are allocated from the data collection node set C of the parking node pi pi Removed;

[0206] C5. Remove the data collection nodes in the remaining UAV flight routes in the memory pool M that are the same as the current UAV flight route; if C pi If it is not empty, return to step C2 until C pi Is empty.

[0207] S233, optimization of drone flight route solution: for the drone flight route solution constructed in step S232, two drone flight routes are randomly selected from the solution, and two data collection nodes with the closest distance are found to exchange and migrate nodes to obtain several new route selection pairs, and the shortest drone flight route is selected from the new route selection pairs to replace the two randomly selected drone flight routes as the optimized drone flight route solution;

[0208] In this scheme, the UAV flight route solution is optimized. After the UAV flight route solution is constructed, the UAV flight route in the solution needs to be optimized. This scheme uses a local search strategy to optimize the UAV flight route; the details are as follows:

[0209] First, randomly select two drone flight routes r1 = {…, c i-1 , c i , c i+1 , ...}, r2={…, c j-1 , c j , c j+1 , ...}, then find the two data collection nodes c that are closest to each other in r1 and r2 i and c j , and finally i and c j Perform the following two operations:

[0210] Data acquisition node exchange (SWAP): exchange c i and c j The positions in r1 and r2; the flight paths of the two exchanged drones are as follows:

[0211] r1 * ={…, c i-1 , c j , c i+1 , …};

[0212] Data collection node migration (move): There are four migration strategies for data collection node migration:

[0213] 1. Change c in r1 i c placed in r2 j Later, the flight routes of the two drones after migration are as follows:

[0214] r1 * ={..., c i-1 , c i+1 , ...};

[0215] 2. Change c in r1 i c placed in r2 j In the previous figure, the flight routes of the two drones after migration are as follows:

[0216] r1 * ={…, c i-1 , c i+1 , …};

[0217] 3. Change c in r2 j c placed in r2 i Later, the flight routes of the two drones after migration are as follows:

[0218] r1 * ={…, c i-1 , c i , c j , c i+1 , …};

[0219] 4. Change C in r2 j c placed in r2 i In the previous figure, the flight routes of the two drones after migration are as follows:

[0220] r1 * ={…, c i-1 , c j , c i , c i+1 , …};

[0221] Furthermore, the steps of optimizing the flight path of the UAV are as follows:

[0222] D1. Solve Rp based on parking node pi and the flight path of the drone at parking node pi. i As input, the optimized UAV flight path is used to solve R pi，new is the output;

[0223] D2. Determine Rp from the flight path of the drone at the parking node pi i Randomly select two UAV flight routes r1 and r2, and find the two data collection nodes c closest to r1 and r2. i and c j ;

[0224] D3, the UAV flight routes r1, r2 and data collection node c i 、c j Perform data collection node exchange and data collection node migration to obtain several new route pairs, and put the several new route pairs and drone flight routes r1 and r2 into the route group;

[0225] D4. Remove the route pairs whose route length exceeds the maximum flight distance of the drone in the route group, and take out the route pair with the longest flight distance and the shortest flight time in the route group, and use this route pair to replace R pi The optimized UAV flight path solution R is obtained by using r1 and r2 in pi，new ; Return to step D2 until the maximum number of iterations is reached.

[0226] S234, memory pool update: put the optimized UAV flight route in the UAV flight route solution into the memory pool, and delete the poor UAV flight route in the memory pool;

[0227] In this solution, for the memory pool update, after the drone flight path solution is optimized, a better solution is obtained; next, the drone flight path in the optimized drone flight path solution will be put into the memory pool, so that a better solution can be constructed in the next iteration process. During the memory pool update process, when the number of drone flight paths in the memory pool reaches the maximum capacity of the memory pool n max When the time comes, the worse drone flight route in the memory pool will be avg (r) Large route deletion. It should be noted that when deleting the poor UAV flight route, it is necessary to check whether the remaining UAV flight routes contain all the data collection nodes at the parking node. If they can be included, the poor route will be deleted directly, otherwise other poor routes will be replaced for deletion.

[0228] S24, for the optimized UAV flight route solution in step S233, the UAV flight route allocation scheme sorts all UAV flight routes and allocates the longest unallocated UAV flight route to the UAV with the shortest total length of the currently allocated route;

[0229] In this scheme, for the UAV flight routes at each parking node that have been solved, it is necessary to assign the UAV flight routes to specific UAVs. For the allocation of UAV flight routes, this scheme adopts the greedy idea to allocate routes. The core idea of ​​this scheme is as follows: First, all UAV flight routes are sorted in descending order according to their length; this step ensures that the algorithm can give priority to allocating longer UAV flight routes; then, each time the algorithm allocates, it will select the UAV with the shortest total length of the currently allocated route, that is, the current The smallest drone is assigned the longest unassigned drone flight route. This not only ensures that each assignment minimizes the current total completion time, but also achieves load balancing.

[0230] Furthermore, the steps of the UAV flight route allocation scheme are as follows:

[0231] F1. Solve R with optimized drone flight path pi，new And the set of drones U that are not assigned with drone flight routes is taken as input to the set of drones U that are assigned with drone flight routes new is the output;

[0232] E2. According to the length l(r) of the UAV flight path, R pi，new Sort the UAV flight routes in descending order;

[0233] E3. Take R pi,new The longest drone flight route in l(r) j , get the time t required for the drone in U to complete the currently assigned task u The smallest drone k , then r j Assign to u k , and updates the total time the drone spends accessing the data collection node at the parking node pi At the same time, r j From R pi，new Removed;

[0234] At this time, if R pi，new If it is not empty, return to step E3 until R pi，new If it is empty, get the set of drones U that assign the flight routes of drones new .

[0235] After completing the above steps, the initial global data collection plan is obtained. Vehicles and drones can collect data according to the initial global data collection plan. However, when new data collection nodes arrive dynamically, a local data collection plan needs to be re-solved.

[0236] The data acquisition scheme adjustment scheme includes: a new data acquisition node division scheme, a local data acquisition solution scheme, and a global data acquisition scheme; the steps for constructing the data acquisition adjustment scheme are as follows:

[0237] S25, the newly added data collection node division scheme defines an unvisited area and an unplanned area, divides the newly added data collection nodes distributed in the unvisited area into an unvisited data collection node set, and divides the newly added data collection nodes distributed in the unplanned area into an unplanned data collection node set;

[0238] Furthermore, the steps of the newly added data collection node division scheme are as follows:

[0239] If a new data collection node is added during the drone's data collection mission, the data collection node needs to be c j is divided, and the steps are as follows:

[0240] F1. Define the area covered by the remaining unvisited parking nodes as the unvisited area Ω uv , the remaining area is the unplanned areaΩ up ;

[0241] F2. Define the unvisited area Ω uv The data collection nodes inside are the unvisited data collection node set C uv , unplanned areaΩ up The data collection nodes inside are the unplanned data collection node set C up ;

[0242] F3. Distribute the newly added data collection nodes in Ω uv With Ω up For the distribution in Ω uv The data collection nodes in C uv For the distribution in Ω up The data collection nodes in C up middle;

[0243] F4, for unvisited areasΩ uv ,UAV flight routes are allocated using a local data collection solution;

[0244] When applied, for C uv The data collection node in C solves a local data collection solution to access. up The data collection nodes in are not considered for the time being.

[0245] S26: For the unvisited data collection node set and the unplanned data collection node set in step S25, there are two situations:

[0246] The first type: the drone is collecting data and the vehicle is parked at the parking node;

[0247] For the first case, the parking node where the vehicle is parked may reach a new data collection node. At this time, it is necessary to re-plan the current flight route of each drone based on the current position and remaining endurance of the drone, and then use the AMP algorithm to solve the flight routes of the drones at the current parking node and the remaining unvisited parking nodes. Finally, the routes are reallocated to each drone to obtain a local data collection plan;

[0248] The second type: all drones are on the vehicle, and the vehicle is moving;

[0249] For the second case, we only need to use the AMP algorithm to solve the flight routes of the remaining unvisited parking nodes, and then reallocate the routes to the drones to obtain a local data collection plan.

[0250] Furthermore, the steps of the local data collection solution are as follows:

[0251] G 1. Take the parking node set P and the data collection node C assigned to each parking node P , the set of drones U that are not assigned with drone flight routes is taken as input, and the set of drones U that are assigned with drone flight routes is taken as input. new is the output;

[0252] G2. If the vehicle is parked at a parking node, obtain the parking node p where the vehicle is parked. now , and p now Delete from P; adjust the current route of the drone based on the current route adjustment plan of the drone, and p based on the solution of the flight route of the drone at the parking node now The UAV flight route at the location is solved; based on the UAV flight route allocation scheme, p now The UAV flight routes at the locations are assigned to the UAVs;

[0253] If the vehicle is traveling at this time, the UAV flight route is solved for each parking node based on the UAV flight route solution at the parking node, and the UAV flight route is allocated to the UAV based on the UAV flight route allocation solution for each parking node.

[0254] Furthermore, when the task arrives dynamically and the drone is performing the task, a new data collection node will be added in the area covered by the parking node where the drone is located. At this time, the drone needs to adjust the current flight route according to the new data collection node to obtain the maximum flight benefit. For the current parking node pi, record E pi is the set of data collection nodes remaining unvisited at pi. This scheme first calculates the current position and remaining endurance of each drone, and then pi Select a data collection node that is closest to all drones, and determine whether the drone can return home smoothly after assigning the data collection node to the drone. If it can, assign it and update the drone position; if not, set the remaining endurance of the drone to 0. Repeat this operation until the endurance of all drones is 0 or the remaining E pi Is empty, thus constructing the current flight routes of all drones.

[0255] Furthermore, the steps of adjusting the current route of the drone are as follows:

[0256] H1, parking node pi, the remaining unvisited data collection node set E at parking node pi pi , the positions and remaining range of all drones are used as input; the adjusted current route of each drone is used as output;

[0257] H2, from E pi Find the nearest drone u from the positions of all drones j and data acquisition node c j ;

[0258] H3. If the drone u j After visiting the data collection node c j After that, the vehicle can be returned smoothly, then c j Join UAV j In the current route, c j From E pi Remove and update the drone j Otherwise, the drone will be j The battery life is set to 0;

[0259] H4. If the endurance of all drones is not 0 and E pi If it is not empty, return to step H2 until all drones or E are updated. pi Is empty.

[0260] S27, during the execution of the local data collection solution in step S26, when the number of data collection nodes in the unplanned data collection node set is greater than the number of data collection nodes in the unvisited data collection node set, the global data collection solution is executed, and the steps are as follows:

[0261] S271, merging the data collection nodes in the unvisited data collection node set and the data collection nodes in the unplanned data collection node set to obtain a global data collection node set;

[0262] S272. Solve the global data collection node set through parking node selection and data collection node allocation scheme to obtain parking nodes and data collection nodes allocated to each parking node; solve the shortest vehicle driving route through the greedy algorithm; for each parking node, obtain the UAV flight route at the parking node through the UAV flight route solution at the parking node; for each parking node, allocate a flight route for the UAV through the UAV flight route allocation scheme.

[0263] In the process of executing the local data collection scheme of this scheme, when the conditions of the global data collection scheme are met, the global data collection scheme must be solved. For the solution of the global data collection scheme, in the process of executing the local data collection scheme, the data collection node set C is not visited. uv The data collection nodes in the will continue to decrease. When the data collection node set C is not planned up The number of data collection nodes in is less than C uv When the number of data collection nodes is , the area Ω is not planned. up The number of distributed data collection nodes is too small, and the time cost of access is high. up The data collection nodes in C are not considered first. up The number of data collection nodes in is greater than or equal to C uv When the number of data collection nodes is medium, the area Ω is not planned at this time up Enough data collection nodes have been distributed. At this time, a global data collection solution needs to be solved to access C uv and C up It is also worth noting that the vehicle route of the global data collection solution is not a closed-loop route, so there is no way to use the ant colony algorithm to solve it. Here, the greedy strategy is also used to solve the vehicle's driving route, that is, each time the nearest unvisited parking node to the current position is selected as the next parking node to stop.

[0264] Furthermore, the steps of the global data collection scheme are as follows:

[0265] K1, collect node set C with unvisited data uv , Unplanned data collection node set C up , the candidate parking node set A is input, and the shortest vehicle driving route r best , assigning drone flight routes to a drone set U new is the output;

[0266] K2, collect the unvisited data node set C uv and the unplanned data collection node set C up The data collection nodes in are merged to obtain the global data collection node set C;

[0267] K3, call the parking node selection and data collection node allocation algorithm to obtain the parking node pi and the data collection node set C assigned to each parking node P ;

[0268] K4, based on the greedy algorithm to obtain the shortest vehicle route r best ;

[0269] K5. For each parking node, based on the solution for the UAV flight route at the parking node, the UAV flight route at the parking node is solved; for each parking node, based on the UAV flight route allocation plan, the UAV set to which the UAV flight route is allocated is obtained.

[0270] Embodiment 2:

[0271] See also Figure 3 , a multi-UAV and vehicle collaborative data acquisition system for dynamic arrival, including:

[0272] A data acquisition model building module 1 is used to build a data acquisition model for collaboration between multiple drones and vehicles, wherein the data acquisition model includes at least one vehicle, several drones and a data center;

[0273] Data collection and adjustment plan building module 2 is used to divide the target area into several node areas and set parking nodes and data collection nodes in each node area; with the goal of minimizing the data collection completion time, an initial global data collection plan and a data collection adjustment plan are built;

[0274] The data collection task execution module 3 is used to execute the data collection task and, when a new task arrives during the data collection task execution, execute the data collection adjustment plan to adjust the current data collection plan;

[0275] The data collection tasks are as follows:

[0276] The vehicle starts from the data center and drives to the first parking node. The drone takes off from the vehicle and drives to the data collection node to collect data. The drone returns to the vehicle and the vehicle drives to the next parking node. This cycle repeats until the vehicle traverses all node areas and returns to the data center, completing the data collection task.

[0277] The initial global data collection scheme includes: parking node selection and data collection node allocation scheme, vehicle route solution scheme, drone flight route solution scheme at parking nodes, and drone flight route allocation scheme;

[0278] Furthermore, for the specific steps of the parking node selection and data collection node allocation scheme, vehicle route solution scheme, drone flight route solution scheme at parking nodes, and drone flight route allocation scheme constructed by the data collection and adjustment scheme construction module 2, please refer to the corresponding parts in Example 1.

[0279] The steps for constructing the initial global data acquisition scheme are as follows:

[0280] S21, based on the parking node selection and data collection node allocation scheme, by calculating the average distance between each candidate parking node and the data collection nodes it covers, the candidate parking nodes with the smallest average distance are selected in turn as parking nodes, and each data collection node is allocated to the nearest parking node, so as to determine the parking nodes and allocate the data collection nodes;

[0281] S22, for the parking nodes determined in step S21, solving the driving route of the vehicle, the vehicle route solving scheme is based on the ant colony algorithm, randomly determining the starting parking node of the ants, continuously obtaining the ant positions and the unvisited parking nodes, calculating the access probability to determine the next parking node, selecting and recording the shortest vehicle route, so as to determine the shortest vehicle driving route of the vehicle;

[0282] S23, for the data collection nodes assigned to each parking node determined in step S21, the flight route of the drone at the parking node is solved by the solution, and the steps are as follows:

[0283] S231, memory pool initialization: putting the flight route of the drone into the memory pool, deleting the data collection nodes included in the flight route of the drone from the data collection node set, and initializing the data collection node set until the maximum capacity of the memory pool is reached;

[0284] S232, UAV flight route solution construction: Use the average access distance to evaluate the UAV flight routes in the memory pool, and select the flight route with the shortest average access distance to construct a solution;

[0285] S233, optimization of drone flight route solution: for the drone flight route solution constructed in step S232, two drone flight routes are randomly selected from the solution, and two data collection nodes with the closest distance are found to exchange and migrate nodes to obtain several new route selection pairs, and the shortest drone flight route is selected from the new route selection pairs to replace the two randomly selected drone flight routes as the optimized drone flight route solution;

[0286] S234, memory pool update: put the optimized UAV flight route in the UAV flight route solution into the memory pool, and delete the poor UAV flight route in the memory pool;

[0287] S24, for the optimized UAV flight route solution in step S233, the UAV flight route allocation scheme sorts all UAV flight routes and allocates the longest unallocated UAV flight route to the UAV with the shortest total length of the currently allocated route;

[0288] The data collection adjustment scheme includes: a new data collection node division scheme, a local data collection solution scheme, and a global data collection scheme;

[0289] Furthermore, for the specific steps of the newly added data collection node division scheme, local data collection solution scheme, and global data collection scheme constructed by the data collection and adjustment scheme construction module 2, please refer to the corresponding parts in Example 1.

[0290] The steps for constructing the data collection and adjustment scheme are as follows:

[0291] S25, the newly added data collection node division scheme divides the newly added data collection nodes distributed in the unvisited area into an unvisited data collection node set and divides the newly added data collection nodes distributed in the unplanned area into an unplanned data collection node set by defining an unvisited area and an unplanned area;

[0292] S26: For the unvisited data collection node set and the unplanned data collection node set in step S25, there are two situations:

[0293] The first type: the drone is collecting data and the vehicle is parked at the parking node; the local data collection solution is solved by solving the drone flight path solution at the parking node, solving the drone flight paths of the current parking node and the remaining unvisited parking nodes, and assigning the flight paths to each drone;

[0294] The second type: all drones are on the vehicle, and the vehicle is driving; then the local data collection solution solves the flight routes of the drones at the remaining unvisited parking nodes through the flight route solution of the drones at the parking nodes, and allocates the flight routes to each drone;

[0295] S27, during the execution of the local data collection solution in step S26, when the number of data collection nodes in the unplanned data collection node set is greater than the number of data collection nodes in the unvisited data collection node set, the global data collection solution is executed, and the steps are as follows:

[0296] S271, merging the data collection nodes in the unvisited data collection node set and the data collection nodes in the unplanned data collection node set to obtain a global data collection node set;

[0297] S272. Solve the global data collection node set through parking node selection and data collection node allocation scheme to obtain parking nodes and data collection nodes allocated to each parking node; solve the shortest vehicle driving route through the greedy algorithm; for each parking node, obtain the UAV flight route at the parking node through the UAV flight route solution at the parking node; for each parking node, allocate a flight route for the UAV through the UAV flight route allocation scheme.

[0298] Embodiment 3:

[0299] See also Figure 4 , In this embodiment, a multi-UAV and vehicle collaborative data acquisition device for dynamic arrival is also included, and the device includes a processor 4 and a memory 5;

[0300] The memory 5 is used to store computer program code 51 and transmit the computer program code 51 to the processor 4;

[0301] The processor 4 is used to execute the multi-UAV and vehicle collaborative data collection method for dynamic arrival described in Example 1 according to the instructions in the computer program code 51.

[0302] In this embodiment, a computer-readable storage medium is also included, in which computer-executable instructions are stored. When the computer-executable instructions are executed on a computer, the multi-UAV and vehicle collaborative data collection method for dynamic arrival described in Example 1 is implemented.

[0303] Generally speaking, the computer instructions for implementing the method of the present invention may be carried in any combination of one or more computer-readable storage media. Non-transitory computer-readable storage media may include any computer-readable media, except for the signal itself that is temporarily propagating.

[0304] The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EKROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, device, or device.

[0305] Computer program code for performing the operation of the present invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, SMalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages, in particular, Python suitable for neural network computing and platform frameworks based on TensorFlow, PyTorch, etc. can be used. The program code can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer or to an external computer (for example, using an Internet service provider to connect via the Internet) through any type of network, including a local area network (LAN) or a wide area network (WAN).

[0306] The above-mentioned device and non-temporary computer-readable storage medium can be found in the specific description of a multi-UAV and vehicle collaborative data collection method for dynamic arrival and its beneficial effects, which will not be repeated here.

[0307] Although the embodiments of the present invention have been shown and described above, it should be understood that the above embodiments are exemplary and should not be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival, characterized in that: include: S1. Constructing a data collection model for multi-UAV and vehicle collaboration, wherein the data collection model includes at least one vehicle, several UAVs and a data center; S2. Divide the target area into several node areas and set parking nodes and data collection nodes in each node area; with the goal of minimizing the data collection completion time, construct an initial global data collection plan and data collection adjustment plan; S3, perform data collection tasks; the data collection tasks are as follows: The vehicle starts from the data center and drives to the first parking node, and then the drone takes off from the vehicle to the data collection node for data collection; the drone returns to the vehicle, and the vehicle drives to the next parking node; This cycle repeats until the vehicle traverses all node areas and returns to the data center, completing the data collection task; The initial global data collection scheme includes: parking node selection and data collection node allocation scheme, vehicle route solution scheme, drone flight route solution scheme at parking nodes, and drone flight route allocation scheme; The steps for constructing the initial global data acquisition scheme are as follows: S21, based on the parking node selection and data collection node allocation scheme, by calculating the average distance between each candidate parking node and the data collection nodes it covers, the candidate parking nodes with the smallest average distance are selected in turn as parking nodes, and each data collection node is allocated to the nearest parking node, so as to determine the parking nodes and allocate the data collection nodes; S22, for the parking nodes determined in step S21, solving the driving route of the vehicle, the vehicle route solving scheme is based on the ant colony algorithm, randomly determining the starting parking node of the ants, continuously obtaining the ant positions and the unvisited parking nodes, calculating the access probability to determine the next parking node, selecting and recording the shortest vehicle route, so as to determine the shortest vehicle driving route of the vehicle; S23, for the data collection nodes assigned to each parking node determined in step S21, the flight route of the drone at the parking node is solved by the solution, and the steps are as follows: S231, memory pool initialization: putting the flight route of the drone into the memory pool, deleting the data collection nodes included in the flight route of the drone from the data collection node set, and initializing the data collection node set until the maximum capacity of the memory pool is reached; S232, UAV flight route solution construction: Use the average access distance to evaluate the UAV flight routes in the memory pool, and select the flight route with the shortest average access distance to construct a route solution; S233, optimization of drone flight route solution: for the drone flight route solution constructed in step S232, two drone flight routes are randomly selected from the solution, and two data collection nodes with the closest distance are found to exchange and migrate nodes to obtain several new route selection pairs, and the shortest drone flight route is selected from the new route selection pairs to replace the two randomly selected drone flight routes as the optimized drone flight route solution; S234, memory pool update: put the optimized UAV flight route in the UAV flight route solution into the memory pool, and delete the poor UAV flight route in the memory pool; S24, for the optimized UAV flight route solution in step S233, the UAV flight route allocation scheme sorts all UAV flight routes and allocates the longest unallocated UAV flight route to the UAV with the shortest total length of the currently allocated route; When a new task arrives during the execution of the above data collection task, the data collection adjustment plan is executed to adjust the current data collection plan; the data collection adjustment plan includes: a new data collection node division plan, a local data collection solution plan, and a global data collection plan; the steps for constructing the data collection adjustment plan are as follows: S25, the newly added data collection node division scheme divides the newly added data collection nodes distributed in the unvisited area into an unvisited data collection node set and divides the newly added data collection nodes distributed in the unplanned area into an unplanned data collection node set by defining an unvisited area and an unplanned area; S26: For the unvisited data collection node set and the unplanned data collection node set in step S25, there are two situations: The first type: the drone is collecting data and the vehicle is parked at the parking node; the local data collection solution is solved by solving the drone flight path solution at the parking node, solving the drone flight paths of the current parking node and the remaining unvisited parking nodes, and assigning the flight paths to each drone; The second type: all drones are on the vehicle, and the vehicle is driving; then the local data collection solution solves the flight routes of the drones at the remaining unvisited parking nodes through the flight route solution of the drones at the parking nodes, and allocates the flight routes to each drone; S27, during the execution of the local data collection solution in step S26, when the number of data collection nodes in the unplanned data collection node set is greater than the number of data collection nodes in the unvisited data collection node set, the global data collection solution is executed, and the steps are as follows: S271, merging the data collection nodes in the unvisited data collection node set and the data collection nodes in the unplanned data collection node set to obtain a global data collection node set; S272. Solve the global data collection node set through parking node selection and data collection node allocation scheme to obtain parking nodes and data collection nodes allocated to each parking node; solve the shortest vehicle driving route through the greedy algorithm; for each parking node, obtain the UAV flight route at the parking node through the UAV flight route solution at the parking node; for each parking node, allocate a flight route for the UAV through the UAV flight route allocation scheme.

2. The method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival according to claim 1 is characterized in that: The steps of the parking node selection and data collection node allocation scheme are as follows: A1, taking the data collection node set C and the candidate parking node set A as input, and taking the parking node set P and the data collection node set C assigned to each parking node P is the output; A2. Calculate the average distance between each candidate parking node and the data collection nodes covered within the maximum flight distance, and sort the parking nodes in ascending order according to the average distance to construct a set of data collection nodes covered by the candidate parking nodes; The expression of the average distance is as follows: in: is the candidate parking node a i The distance from the data collection node cj; Then for candidate parking node a i The covered data collection node set C ai It is expressed as follows: Where: d uav is the maximum flight distance of the drone; A3. Select the candidate parking node with the smallest average distance as the parking node, remove the candidate parking node from the candidate parking node set A; remove the data collection node covered by the candidate parking node from the data collection node set C; add the candidate parking node to the parking node set P; repeat the above operation until the data collection node set C is empty; A4. For each data collection node, find the parking node closest to it and assign the data collection node to the parking node; the expression is as follows: C pi =C pi ∪{c i }; in: C pi is the data collection node assigned to parking node pi, and ci is the i-th data collection node in C.

3. The method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival according to claim 1 is characterized in that: The steps of the vehicle routing solution are as follows: B1, based on the vehicle route node set r truck =P∪n0 as input, with the shortest vehicle route r best is the output; where: n0 is the data center node; B2. Randomly generate the starting parking node for each ant and record it in the vehicle route table path_table; B3. Get the parking node n where the ant is currently located visting And the unvisited parking node list unvisited_list; B4. Calculate the access probability of each unvisited parking node and determine the next parking node n to be visited. next ; B5. Update the vehicle route table path_table, the unvisited parking node list unvisited_list and the current ant location n visting ←n next ; If there are still parking nodes that the current ant has not visited, return to step B4; if there are still untraversed ants, return to step B3; if all ants have been traversed and there are no unvisited parking nodes for the current ant, proceed to step B6; B6. Get the shortest vehicle route r in the vehicle route table path_table shortest , and record the shortest vehicle route R in this iteration shortest =R shortest ∪{r shortest }, then calculate the pheromone increment on the vehicle route and update the pheromone matrix M phe ; B7. Iterate according to steps B2-B6 until the maximum number of iterations is reached, and the shortest vehicle route r shortest Get the shortest vehicle route r best .

4. The method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival according to claim 1 is characterized in that: The steps of the UAV flight path solution construction are as follows: C1, memory pool M, parking node pi and data collection node set allocated by parking node pi As input, the flight path of the UAV at the parking node pi is the output; C2. Sort the UAV flight routes in the memory pool in ascending order according to the average access distance of the UAV flight routes; the expression of the average access distance is as follows: Where; l(r) is the length of the UAV flight route, |r| is the number of data collection nodes in the UAV flight route; C3. Select a drone flight route with the shortest average access distance and add it to and removed from memory pool M; C4, the data collection nodes in the current UAV's flight path are allocated from the data collection node set of parking node pi Removed; C5. Remove the data collection nodes in the remaining UAV flight routes in the memory pool M that are the same as the current UAV flight route; if is not empty, return to step C2 until Is empty.

5. The method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival according to claim 1 is characterized in that: The steps of optimizing the UAV flight path are as follows: D1, based on parking node pi, the UAV flight path solution at parking node pi As input, the optimized UAV flight path solution is the output; D2. The flight path of the drone is solved from the parking node pi. Randomly select two UAV flight routes r1 and r2, and find the two data collection nodes c closest to r1 and r2. i and c j ; D3, the UAV flight routes r1, r2 and data collection node c i 、c j Perform data collection node exchange and data collection node migration to obtain several new route pairs, and put the several new route pairs and drone flight routes r1 and r2 into the route group; D4. Remove the route pairs in the route group whose route length exceeds the maximum flight distance of the drone, and take out the route pair with the longest flight distance and the shortest flight time in the route group, and use this route pair to replace The optimized UAV flight path solution is obtained by using r1 and r2 in Return to step D2 until the maximum number of iterations is reached.

6. The method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival according to claim 5 is characterized in that: The steps of the UAV flight route allocation scheme are as follows: F1, solve the problem with the optimized UAV flight path And the set of drones U that are not assigned with drone flight routes is taken as input to the set of drones U that are assigned with drone flight routes new is the output; E2. According to the length l(r) of the UAV flight route Sort the UAV flight routes in descending order; E3. Take The longest drone flight route in l(r) j , take the time t required for the drone in U to complete the currently assigned task u The smallest drone k , then r j Assign to u k , and updates the total time the drone spends accessing the data collection node at the parking node pi At the same time, r j from Removed; If at this time If it is not empty, return to step E3 until If it is empty, get the set of drones U that assign the flight routes of drones new .

7. The method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival according to claim 6 is characterized in that: The steps of the newly added data collection node division scheme are as follows: If a new data collection node is added during the UAV's data collection mission, the data collection node cj needs to be divided at this time. The steps are as follows: F1. Define the area covered by the remaining unvisited parking nodes as the unvisited area Ω uv , the remaining area is the unplanned areaΩ up ; F2. Define the unvisited area Ω uv The data collection nodes inside are the unvisited data collection node set C uv , unplanned areaΩ up The data collection nodes inside are the unplanned data collection node set C up ; F3. Distribute the newly added data collection nodes in Ω uv With Ω up For the distribution in Ω uv The data collection nodes in C uv For the distribution in Ω up The data collection nodes in C up middle; F4, for unvisited areasΩ uv ,UAV flight routes are allocated using a local data collection solution; The steps of the local data collection solution are as follows: G1, parking node set P, data collection node C assigned to each parking node P , the set of drones U that are not assigned with drone flight routes is taken as input, and the set of drones U that are assigned with drone flight routes is taken as input. new is the output; G2. If the vehicle is parked at a parking node, obtain the parking node p where the vehicle is parked. now , and p now Delete from P; adjust the current route of the drone based on the current route adjustment plan of the drone, and p based on the solution of the flight route of the drone at the parking node now Solve the flight path of the UAV at the location; Based on the UAV flight route allocation scheme, p now The UAV flight routes at the locations are assigned to the UAVs; If the vehicle is traveling at this time, the UAV flight route is solved for each parking node based on the UAV flight route solution at the parking node, and the UAV flight route is allocated to the UAV based on the UAV flight route allocation solution for each parking node.

8. The method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival according to claim 7 is characterized in that: The steps of the current route adjustment plan of the drone are as follows: H1, the parking node pi, the remaining unvisited data collection node set at the parking node pi The input is the position and remaining range of all drones; the output is the adjusted current route of each drone; H2. From Find the nearest drone u from the positions of all drones j and data collection node cj; H3. If the drone u j After visiting the data collection node cj, you can return to the vehicle smoothly, then add cj to the drone u j In the current route, cj is changed from Remove and update the drone j Otherwise, the drone will be j The battery life is set to 0; H4. If the endurance of all drones is not 0 and If it is not empty, return to step H2 until all drones are updated or Is empty.

9. The method for collaborative data collection of multiple UAVs and vehicles for dynamic arrival according to claim 1, characterized in that: The steps of the global data acquisition scheme are as follows: K1, collect node set C with unvisited data uv , Unplanned data collection node set C up , the candidate parking node set A is input, and the shortest vehicle driving route r best , assigning drone flight routes to a drone set U new is the output; K2, collect the unvisited data node set C uv and the unplanned data collection node set C up The data collection nodes in are merged to obtain the global data collection node set C; K3, call the parking node selection and data collection node allocation algorithm to obtain the parking node pi and the data collection node set C assigned to each parking node P ; K4, based on the greedy algorithm to obtain the shortest vehicle route r best ; K5. For each parking node, a solution is obtained based on the flight route of the UAV at the parking node to obtain the flight route of the UAV at the parking node; For each parking node, based on the UAV flight route allocation scheme, a set of UAVs with assigned UAV flight routes is obtained.

10. A multi-UAV and vehicle collaborative data acquisition system for dynamic arrival, characterized by: The system is applied to the method described in any one of claims 1 to 9, and the system comprises: A data collection model building module (1) is used to build a data collection model for cooperation between multiple drones and vehicles, wherein the data collection model includes at least one vehicle, several drones and a data center; The data collection and adjustment scheme building module (2) is used to divide the target area into a number of node areas and set a parking node and a data collection node in each node area; with the goal of minimizing the data collection completion time, an initial global data collection scheme and a data collection adjustment scheme are built; A data collection task execution module (3) is used to execute the data collection task and, when a new task arrives during the execution of the data collection task, execute the data collection adjustment plan to adjust the current data collection plan; The data collection tasks are as follows: The vehicle starts from the data center and drives to the first parking node. The drone takes off from the vehicle and drives to the data collection node to collect data. The drone returns to the vehicle and the vehicle drives to the next parking node. This cycle repeats until the vehicle traverses all node areas and returns to the data center, completing the data collection task. The initial global data collection scheme includes: parking node selection and data collection node allocation scheme, vehicle route solution scheme, drone flight route solution scheme at parking nodes, and drone flight route allocation scheme; The steps for constructing the initial global data acquisition scheme are as follows: S21, based on the parking node selection and data collection node allocation scheme, by calculating the average distance between each candidate parking node and the data collection nodes it covers, the candidate parking nodes with the smallest average distance are selected in turn as parking nodes, and each data collection node is allocated to the nearest parking node, so as to determine the parking nodes and allocate the data collection nodes; S22, for the parking nodes determined in step S21, solving the driving route of the vehicle, the vehicle route solving scheme is based on the ant colony algorithm, randomly determining the starting parking node of the ants, continuously obtaining the ant positions and the unvisited parking nodes, calculating the access probability to determine the next parking node, selecting and recording the shortest vehicle route, so as to determine the shortest vehicle driving route of the vehicle; S23, for the data collection nodes assigned to each parking node determined in step S21, the flight route of the drone at the parking node is solved by the solution, and the steps are as follows: S231, memory pool initialization: putting the flight route of the drone into the memory pool, deleting the data collection nodes included in the flight route of the drone from the data collection node set, and initializing the data collection node set until the maximum capacity of the memory pool is reached; S232, UAV flight route solution construction: Use the average access distance to evaluate the UAV flight routes in the memory pool, and select the flight route with the shortest average access distance to construct a solution; S233, optimization of drone flight route solution: for the drone flight route solution constructed in step S232, two drone flight routes are randomly selected from the solution, and two data collection nodes with the closest distance are found to exchange and migrate nodes to obtain several new route selection pairs, and the shortest drone flight route is selected from the new route selection pairs to replace the two randomly selected drone flight routes as the optimized drone flight route solution; S234, memory pool update: put the optimized UAV flight route in the UAV flight route solution into the memory pool, and delete the poor UAV flight route in the memory pool; S24, for the optimized UAV flight route solution in step S233, the UAV flight route allocation scheme sorts all UAV flight routes and allocates the longest unallocated UAV flight route to the UAV with the shortest total length of the currently allocated route; The data collection adjustment scheme includes: a new data collection node division scheme, a local data collection solution scheme, and a global data collection scheme; The steps for constructing the data collection and adjustment scheme are as follows: S25, the newly added data collection node division scheme divides the newly added data collection nodes distributed in the unvisited area into an unvisited data collection node set and divides the newly added data collection nodes distributed in the unplanned area into an unplanned data collection node set by defining an unvisited area and an unplanned area; S26: For the unvisited data collection node set and the unplanned data collection node set in step S25, there are two situations: The first type: the drone is collecting data and the vehicle is parked at the parking node; the local data collection solution is solved by solving the drone flight path solution at the parking node, solving the drone flight paths of the current parking node and the remaining unvisited parking nodes, and assigning the flight paths to each drone; The second type: all drones are on the vehicle, and the vehicle is driving; then the local data collection solution solves the flight routes of the drones at the remaining unvisited parking nodes through the flight route solution of the drones at the parking nodes, and allocates the flight routes to each drone; S27, during the execution of the local data collection solution in step S26, when the number of data collection nodes in the unplanned data collection node set is greater than the number of data collection nodes in the unvisited data collection node set, the global data collection solution is executed, and the steps are as follows: S271, merging the data collection nodes in the unvisited data collection node set and the data collection nodes in the unplanned data collection node set to obtain a global data collection node set; S272. Solve the global data collection node set through parking node selection and data collection node allocation scheme to obtain parking nodes and data collection nodes allocated to each parking node; solve the shortest vehicle driving route through the greedy algorithm; for each parking node, obtain the UAV flight route at the parking node through the UAV flight route solution at the parking node; for each parking node, allocate a flight route for the UAV through the UAV flight route allocation scheme.

Images