Network communication task planning method and device based on multiple unmanned aerial vehicles, equipment and medium

Abstract

The invention relates to a network communication task planning method and device based on multiple unmanned aerial vehicles, equipment and a medium, and the method comprises the steps: finding a group of communication device node setting positions enabling the communication area to be maximum in a to-be-recovered communication region, and enabling the group of positions to serve as target positions of a multi-unmanned aerial vehicle launching task; taking the shortest total flight path of each unmanned aerial vehicle as a target, and taking the initial position of each unmanned aerial vehicle as a clustering center to cluster the multi-unmanned aerial vehicle launching tasks to obtain a launching task set corresponding to each unmanned aerial vehicle; and the optimized marine predator algorithm is adopted to plan the paths of all the delivery tasks in the corresponding delivery task set completed by each unmanned aerial vehicle, and network communication task planning of each unmanned aerial vehicle is obtained. By adopting the method, the communication of most areas in the communication area to be recovered can be effectively and quickly recovered.

Description

Technical Field

[0001] The present application relates to the technical field of unmanned aerial vehicle task planning, and in particular to a method, device, equipment and medium for network communication task planning based on multiple unmanned aerial vehicles. Background Art

[0002] When an earthquake or fire occurs, it may cause a problem that the network in a small area cannot be connected. In this case, a certain number of wireless network connectors can be deployed by drones to maximize the connection of these small areas, such as Figure 1 As shown. The wireless network connector is equivalent to a mobile base station, which can be used for satellite communication and ground communication and can be quickly deployed and moved. Due to its strong mobility and adaptability to various environments, drones can be used as aerial platforms and carry multiple network communication equipment, playing an important role in urban rescue and post-disaster construction.

[0003] However, there is currently no unified standard or mature deployment plan for drones to implement communication tasks. There are still many technical difficulties in network coverage optimization, equipment resource scheduling, and multi-drone route planning in dynamic environments. Summary of the invention

[0004] Based on this, it is necessary to provide a network communication task planning method, device, equipment and medium based on multiple UAVs that can quickly solve the problem of network connectivity in small areas in response to the above technical problems.

[0005] A network communication task planning method based on multiple unmanned aerial vehicles, the method comprising:

[0006] Obtain the area of ​​the connection area to be restored, the number of connector nodes, and the initial positions of multiple drones;

[0007] Find a group of connector node settings in the area to be restored that maximizes the connectivity area, and use the group of locations as the target locations for the multi-UAV delivery mission;

[0008] With the goal of minimizing the total flight path of each drone, the initial position of each drone is used as a clustering center to cluster the multiple drone delivery tasks to obtain a delivery task set corresponding to each drone;

[0009] For each of the delivery task sets, the optimized ocean predator algorithm is used to plan the paths for each drone to complete all delivery tasks in the corresponding delivery task set, so as to obtain the network communication task plan for each drone.

[0010] In one embodiment, finding a set of connector node locations that maximize the connectivity area in the connectivity area to be restored includes:

[0011] Dividing the area to be restored into a plurality of circular sub-areas with different areas, each of the sub-areas is represented on a two-dimensional plane as the coordinates of the center point and the radius of the circle;

[0012] Obtaining the connected area of ​​each of the connector nodes according to the position coordinates and the sensing radius of each of the connector nodes;

[0013] Determine whether the sub-area is covered by the connector node according to the distance from the center of the sub-area to the center of the connector node and the connector sensing radius;

[0014] The regional connectivity rate of all connector nodes in the connectivity area to be restored is defined as the ratio of the unconnected sub-area covered by the connector node set to the total area of ​​the connectivity area to be restored;

[0015] Find a set of connector node coordinates that maximizes the connectivity rate of the area.

[0016] In one embodiment, a 0-1 perception model is used to obtain a set of connector node coordinates that maximizes the connectivity rate of the area.

[0017] In one embodiment, when clustering the delivery tasks, a k-means clustering algorithm is used.

[0018] In one embodiment, when the optimized marine predator algorithm is used for path planning for each drone, the objective function is expressed as:

[0019]

[0020] In the above formula, the set of delivery tasks corresponding to a certain drone is expressed as V = {v1,v2,v3,...,v A}, the order of placing the connector nodes is expressed as T = {t1, t2, t3, ..., t A}, where t i =V(i=1,...,A), and t i+1 =t1.

[0021] In one embodiment, the optimized marine predator algorithm is combined with the ε-greedy strategy on the basis of the marine predator algorithm, and the strategy to be adopted is determined according to a preset ε value and a random number in different iteration stages.

[0022] In one embodiment, the optimized marine predator algorithm includes:

[0023] In the initial stage of iteration, if the random number is less than the preset ε value, the first strategy is used to search in a large range. If the random number is greater than or equal to the preset ε value, the second strategy or the third strategy is randomly used to search.

[0024] In the middle stage of the iteration, if the random number is less than the preset ε value, the second strategy is randomly adopted to search in a small range; if the random number is greater than or equal to the preset ε value, the first strategy or the third strategy is randomly adopted to search;

[0025] At the end of the iteration, if the random number is less than the preset ε value, the third strategy is randomly adopted to search in the local range; if the random number is greater than or equal to the preset ε value, the first strategy or the second strategy is randomly adopted to search.

[0026] The present application also provides a network communication task planning device based on multiple UAVs, the device comprising:

[0027] The communication task information acquisition module is used to obtain the area of ​​the connection area to be restored, the number of connector nodes, and the initial positions of multiple drones;

[0028] A delivery task location obtaining module is used to find a set of connector node setting locations that maximize the connectivity area in the connectivity area to be restored, and use the set of locations as the target locations for the multi-UAV delivery task;

[0029] The UAV task allocation module is used to cluster the multiple UAV delivery tasks with the shortest total flight path of each UAV as the goal, using the initial position of each UAV as the clustering center, and obtaining a delivery task set corresponding to each UAV;

[0030] The UAV task route planning module is used to plan the path for each UAV to complete all delivery tasks in the corresponding delivery task set using the optimized marine predator algorithm for each UAV to obtain the network communication task plan for each UAV.

[0031] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0032] Obtain the area of ​​the connection area to be restored, the number of connector nodes, and the initial positions of multiple drones;

[0033] Find a group of connector node settings in the area to be restored that maximizes the connectivity area, and use the group of locations as the target locations for the multi-UAV delivery mission;

[0034] With the goal of minimizing the total flight path of each drone, the initial position of each drone is used as a clustering center to cluster the multiple drone delivery tasks to obtain a delivery task set corresponding to each drone;

[0035] For each of the delivery task sets, the optimized ocean predator algorithm is used to plan the paths for each drone to complete all delivery tasks in the corresponding delivery task set, so as to obtain the network communication task plan for each drone.

[0036] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0037] Obtain the area of ​​the connection area to be restored, the number of connector nodes, and the initial positions of multiple drones;

[0038] Find a group of connector node settings in the area to be restored that maximizes the connectivity area, and use the group of locations as the target locations for the multi-UAV delivery mission;

[0039] With the goal of minimizing the total flight path of each drone, the initial position of each drone is used as a clustering center to cluster the multiple drone delivery tasks to obtain a delivery task set corresponding to each drone;

[0040] For each of the delivery task sets, the optimized ocean predator algorithm is used to plan the paths for each drone to complete all delivery tasks in the corresponding delivery task set, so as to obtain the network communication task plan for each drone.

[0041] The above-mentioned network communication task planning method, device, equipment and medium based on multiple drones, by finding a set of connector node setting positions that maximize the connected area in the Unicom area to be restored, and using this set of positions as the target position of the multi-drone delivery task, with the shortest total flight path of each drone as the goal, clustering the multi-drone delivery tasks with the initial position of each drone as the clustering center, and obtaining a delivery task set corresponding to each drone, and then for each delivery task set, using the optimized marine predator algorithm to plan the path for each drone to complete all delivery tasks in the corresponding delivery task set, and obtaining the network communication task planning of each drone. This method can effectively and quickly restore the communication of most areas in the Unicom area to be restored. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic diagram of a local communication network connection in one embodiment;

[0043] Figure 2 It is a flowchart of a method for planning a network communication task based on multiple UAVs in one embodiment;

[0044] Figure 3 A schematic diagram of a flow chart of the steps of planning a network communication task based on multiple UAVs in one embodiment;

[0045] Figure 4 A schematic diagram of regional connectivity in an embodiment;

[0046] Figure 5 This is a schematic diagram of the location distribution of unconnected sub-areas in the area to be restored in an experiment;

[0047] Figure 6 A schematic diagram of the locations of sub-areas and connectors in the interconnection area to be restored in an experiment;

[0048] Figure 7 A schematic diagram of the clustering results of the UAV launch task in an experiment;

[0049] Figure 8 A schematic diagram of the path planning for each drone to deploy multiple wireless connectors in scenarios 1 and 2 of an experiment;

[0050] Fig. 9 It is a structural block diagram of a network communication task planning device based on multiple UAVs in one embodiment;

[0051] Fig.10 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0053] In this application, if Figure 2 As shown, a network communication task planning method based on multiple UAVs is provided, which specifically includes the following steps:

[0054] Step S100, obtaining the area of ​​the connection area to be restored, the number of connector nodes, and the initial positions of multiple drones.

[0055] Step S110, finding a group of connector node setting positions that maximize the connectivity area in the connectivity area to be restored, and using the group of positions as the target positions for the multi-UAV delivery mission.

[0056] Step S120 , with the goal of minimizing the total flight path of each drone, clustering the multiple drone delivery tasks using the initial position of each drone as a clustering center to obtain a delivery task set corresponding to each drone.

[0057] Step S130, for each delivery task set, the optimized ocean predator algorithm is used to plan the path for each drone to complete all delivery tasks in the corresponding delivery task set, and obtain the network communication task plan for each drone.

[0058] In this application, an improved marine predator algorithm is proposed and combined with a clustering algorithm to deploy wireless network connectors using multiple UAVs, and the improved algorithm is used for wireless network connector position estimation and multi-UAV motion planning.

[0059] After the problem of wireless network connector deployment is handled by multiple UAVs, it is converted into multi-UAV task planning, which includes two aspects: task allocation and task path planning.

[0060] Furthermore, the communication network recovery model is constructed into three parts, namely, wireless network connector location estimation in step S110, drone multi-tasking processing in step S120, and drone motion planning in step S130.

[0061] like Figure 3 As shown in the figure, the overall process of the multi-UAV network connection task is given. Since in real environments, the wireless connector is often limited in its capabilities when facing sudden communication failures, it is necessary to select the location of the wireless connector according to the location of the communication interruption area to maximize the effectiveness of the wireless connector. Since multiple drones can be used for delivery, the tasks can be divided by clustering, and then the movement route is planned according to the tasks assigned to each drone. For the wireless connector position estimation problem, the input is the location of the communication interruption area, and for the multi-UAV task clustering problem, the input is the performance parameters of the wireless connector, and the output is the placement location of the wireless connector. For the multi-UAV task clustering problem, the input is the estimated location of the wireless connector and the number of available drones, and the output is the number group of wireless connectors that each drone needs to drop. For the multi-UAV motion planning problem, the input is the number group of wireless connectors dropped by each drone and the starting position information of the drone, and the output is the order and flight route of the wireless connectors dropped by each drone.

[0062] In step S100, first obtain the area of ​​the unconnected area to be restored, the number of connectors to be deployed and the coverage radius of each connector, the perception radius, the number of drones involved in deploying the connectors and the initial position of each drone.

[0063] In step S110, finding a set of connector node setting positions that maximize the connectivity area in the connectivity area to be restored includes: first dividing the connectivity area to be restored into multiple circular sub-areas with different areas, each sub-area is represented on a two-dimensional plane as the coordinates of the center point of the circle and a radius, obtaining the connectivity area of ​​each connector node according to the position coordinates and the perception radius of each connector node, and defining whether the sub-area is covered by the connector node according to the distance from the center of the sub-area to the center of the connector node and the perception radius of the connector, defining the regional connectivity rate of all connector nodes in the connectivity area to be restored as the ratio of the unconnected sub-area covered by the connector node set to the total area of ​​the connectivity area to be restored, and finding a set of connector node coordinates that maximize the regional connectivity rate.

[0064] In this embodiment, a 0-1 perception model is used to obtain a set of node coordinates of the connector that maximizes the regional connectivity rate. In other embodiments, the connectors may be placed using a grid method, where all areas are divided into grids of equal size, and the grids are randomly or sequentially selected for placement of the connectors.

[0065] Specifically, the regional communication method is through the regional wireless connector. When the area is within the communication range of the wireless connector, the area can be regarded as a communication area, such as Figure 4 shown.

[0066] Furthermore, the connection range of the connector node on the two-dimensional plane is a circle with the node as the center and a radius of R. n The circular area, R n It is called the connection radius. The connection radius is related to the physical characteristics of the built-in components of the node. Assume that the position coordinates of the connector node n are (x n ,y n ), then in the 0-1 perception model, for any point p(x p ,y p ), then the probability of the event occurring when the connector node n monitors the point p in the region is expressed as:

[0068]

[0069] In formula (1), is the Euclidean distance between points and .

[0070] Assume that the number of wireless connector nodes with the same structure deployed in the unconnected area is N, the position coordinate value of each connector node is assumed to have been initialized, and the perception radius of the connector node is r. Then the entire connector node set can be expressed as:

[0071] Node{x1,...,x N}(2)

[0072] In formula (2), node i ={x i ,y i ,r}, indicating that the node (x i ,y i ) is the center of the circle, and r is the sensing radius. At the same time, assuming that the entire area to be restored has n circular sub-areas, that is, the entire area to be restored is divided into multiple circular sub-areas, and the coordinates of the center point of each sub-area are expressed as (x, y). Then the distance between the center point of the sub-area and the center of a certain connector node is expressed as:

[0073]

[0074] The event that the unconnected sub-area is connected by the connector node is defined as c i . Then the event occurs when the target point (x, y) is detected by the sensor node node i cover:

[0075]

[0076] Next, the regional connectivity rate of all connector nodes in the unconnected sub-area is defined as: the ratio of the unconnected sub-area covered by the connector node set to the total unconnected area, i.e., the area of ​​the area to be restored, as shown in the formula:

[0077]

[0078] In this way, we only need to find a set of connector node positions that maximize the area of ​​the connected area to be restored. This set of connector node positions is used as the location of the drone delivery task, and also generates the tasks to be planned for subsequent drone planning.

[0079] In step S120, when multiple drones jointly complete the network restoration task, they start from different directions to make the total flight path of the drones shortest. Since the number of drones is specified in advance, clustering is used to cluster the delivery tasks of the connectors to obtain the results of the delivery tasks that each drone needs to complete. Partition clustering attempts to divide the data set X into K clusters C = {C1, C2, ..., C K}, so that it satisfies the following conditions: i,j=1,...,K i≠j. That is, the initial position of each UAV is used as the clustering center, the delivery tasks, i.e., the deployment positions of each connector node, are clustered, and the clusters obtained after clustering are assigned to the corresponding UAVs for execution.

[0080] In this embodiment, when clustering the delivery tasks, the k-means clustering algorithm is used.

[0081] Specifically, clustering divides a large amount of data into categories with different characteristics. The K-means clustering step is a cyclic iterative algorithm. Suppose you want to cluster N samples and require them to be clustered into K categories. First, select K points as the initial center points. Then, divide all observations into the class where each center point is located. There are several observations in each class. Calculate the average value of all sample points in the K class as the K center points for the second iteration. Then, repeat steps 2 and 3 until convergence (the center point no longer changes or reaches the specified number of iterations), and the clustering process ends.

[0082] Furthermore, given a sample set X = {x1,...,x n The goal of the K-means clustering algorithm is to minimize the mean square error of the clusters obtained after clustering. The mean square error (MSE) can be expressed as

[0083]

[0084] In formula (6), u i Represents the cluster C i Reducing the mean square error loss makes the vectors in the cluster more compact, which can achieve better classification results.

[0085] In step S130, when planning the drone path, certain rules are set to allow the drone to place these wireless connectors in order, and the order is obtained by using a swarm intelligence algorithm. The existing standard description of the TSP (Traveling Salesman Problem) problem is: it is known that there are A wireless sensor nodes, and a drone starts from one of the nodes and needs to place all the wireless connectors on the way, but the number of times it passes is one and only one, and finally returns to the starting area node. How to plan the route so that the drone takes the shortest route.

[0086] Suppose the set of delivery missions of a certain drone is V = {v1,v2,v3,...,v A}, the order of placing the connector nodes is T = {t1, t2, t3, ..., t A}, where t i =V(i=1,...,A), and t i+1 =t1, then the objective function of the problem is as follows:

[0087]

[0088] Then the optimal value of the objective function is the shortest path and time required to deploy all wireless connector drones.

[0089] In this embodiment, the optimized marine predator algorithm is combined with the ε-greedy strategy on the basis of the marine predator algorithm. In different iteration stages, the strategy to be adopted is determined according to a preset ε value and a random number.

[0090] Specifically, the marine predator algorithm (MPA) is a meta-heuristic algorithm inspired by the predation process of marine organisms. Since the marine predator chooses the movement strategy in Lévy and Brown, it has good optimization performance. The theoretical basis of the marine predator algorithm mainly involves the behavior and ecology of marine organisms, swarm intelligence, and meta-heuristic thinking. By incorporating these theoretical principles into the algorithm design, the algorithm can simulate the optimization strategy of marine predators and apply it to optimization and search to solve practical problems.

[0091] Furthermore, the MPA algorithm uses random initialization of prey positions within the search space to start the optimization process. The MPA algorithm includes multiple different iterative stages. The mathematical description is as follows:

[0092] X0=X min +rand(X max -X min ) (8)

[0093] At the beginning of the iteration, when the predator speed is faster than the prey speed, the mathematical description of the MPA optimization process based on the exploration strategy is as shown in formula (9), which can be regarded as strategy A1.

[0094]

[0095] In formula (9), stepsize i is the moving step length, R B is a normally distributed Brownian walk random vector, Elite i An elite matrix constructed by top predators, Prey i is the prey matrix with the same dimension as the elite matrix, is a term-by-term multiplication operator, P is equal to 0.5, R is a uniform random vector in [0, 1], n is the population size, iter and max_iter are the current and maximum number of iterations respectively.

[0096] In the middle of the iteration, when the speed of the predator and the prey is the same, the prey is responsible for exploitation based on the Lévy walk strategy; the predator is responsible for exploration based on the Brownian walk strategy, and gradually switches from the exploration strategy to the exploitation strategy. The mathematical description of exploitation and exploration is shown in formulas (10) and (11), which can be regarded as strategy A2.

[0097]

[0098] In formula (10) and formula (11), RL is a random vector with Levy distribution; CF = (1-iter / max_iter) (2iter / max_iter) , CF is an adaptive parameter that controls the predator's moving step size.

[0099] At the end of the iteration, when the predator is slower than the prey, the predator adopts an exploitation strategy based on the Lévy walk. Its mathematical description is shown in formula (12), which can be regarded as strategy A3.

[0100]

[0101] In formula (12), the FADs effect or eddy current is taken into account. Fish aggregation devices (FADs) or eddy current effects usually change the foraging behavior of marine predators. This strategy enables MPA to overcome the premature convergence problem and escape the local extreme value during the optimization process. Its mathematical description is as follows:

[0102]

[0103] In formula (13), FADs is the impact probability, and 0.2 is selected as the priority. U is a binary vector, r is a random number in [0,1], and r1 and r2 are random indices of the prey matrix, respectively.

[0104] That is to say, in the existing marine predator algorithm, the strategy adopted is fixed at different iteration stages. This may include: 1. Slow convergence. The fixed strategy may cause the algorithm to move slowly in the search space and fail to effectively explore potential high-quality solutions. Therefore, more iterations may be required to find a satisfactory solution. 2. Falling into the local optimum. Since the strategy remains unchanged, the algorithm may linger near the local optimal solution and cannot jump out, missing the opportunity to find the global optimal solution. This is particularly evident in complex optimization problems. 3. Insufficient adaptability. The fixed strategy lacks the ability to dynamically adjust the problem characteristics or the current search status. Different search strategies may be required at different stages to adapt to the complexity and variability of the problem. 4. Reduced diversity. During the search process, the fixed strategy may lead to a decrease in group diversity, making the search process convergent and reducing the breadth of the algorithm's exploration in the search space. 5. Premature convergence. If the algorithm determines the strategy too early at a certain stage, it may lead to a narrowing of the search scope, thereby reducing the overall performance of the algorithm.

[0105] In this method, the ε-greedy strategy is selected to improve the marine predator algorithm. The ε-greedy algorithm is an algorithm that balances exploration and exploitation and is a commonly used strategy selection method in reinforcement learning. Under the theoretical framework of reinforcement learning, the ε-greedy strategy, as a behavioral strategy, provides a simple and intuitive method for agents to select actions in the environment.

[0106] Specifically, the optimized marine predator algorithm includes: in the initial stage of iteration, if the random number is less than the preset ε value, the first strategy (the above-mentioned strategy A1) is used to search in a large range, and if the random number is greater than or equal to the preset ε value, the second strategy (the above-mentioned strategy A2) or the third strategy (the above-mentioned strategy A3) is randomly used to search. In the middle stage of iteration, if the random number is less than the preset ε value, the second strategy is randomly used to search in a small range, and if the random number is greater than or equal to the preset ε value, the first strategy or the third strategy is randomly used to search. In the final stage of iteration, if the random number is less than the preset ε value, the third strategy is randomly used to search in a local range, and if the random number is greater than or equal to the preset ε value, the first strategy or the second strategy is randomly used to search.

[0107] In this embodiment, a pseudo code for implementing the above-mentioned optimized marine predator algorithm is also provided, as shown below:

[0108]

[0109]

[0110] Furthermore, in this embodiment, a pseudo code for implementing steps S120 to S130 is also proposed, which is expressed as:

[0111]

[0112]

[0113] In this paper, the effectiveness of the proposed method is also demonstrated through simulation experimental results.

[0114] In the simulation experiment, the area to be connected is set to be 500km long and 500km wide. The number of drones used to launch wireless connectors is 6, the number of connectors is 50, and the signal radius is 15km. The maximum radius of the connection sub-area is 30km and the minimum radius is 5km. The theoretical communication range of the GSM base station is tens of kilometers. Considering the obstruction of intermediate buildings and different types of equipment, it is set to 15km. Due to the large area, long-range drones should be selected, such as Global Eagle, Heron, Orion, etc. The position of the sub-area to be connected is randomly generated in the spatial area, such as Figure 5 shown.

[0115] The motion constraints of the UAV system are mainly reflected in dynamic constraints, collision avoidance constraints, energy constraints, etc. Tasks are not planned at the campaign level, but at the tactical level. The scope of this experiment is relatively wide, and it mainly considers the task allocation and path planning of UAVs. Since the speed and model of the UAV are not restricted, the meteorological conditions during navigation are unknown, and the impact of the dynamic constraints of the UAV system such as turning radius and pitch angle on the total range is not considered. For the collision avoidance problem of the UAV swarm, this paper adopts the method of setting different flight altitudes to avoid collisions during movement. For the endurance problem of the UAV, a UAV model with an endurance exceeding the total range calculated by the experiment and maintaining a certain safety margin should be used.

[0116] The experiment used five swarm intelligence optimization algorithms for position estimation, including particle swarm optimization (PSO), Harris Eagle optimization algorithm (HHO), slime mold algorithm (SMA), marine predator algorithm (MPA) and ε-marine predator algorithm (ε-MPA), which is the optimized marine predator algorithm proposed in this paper. These algorithms include classic and newly emerged swarm intelligence algorithms in recent years. The algorithm parameter settings are shown in Table 1, where the ε-MPA and MPA algorithm parameter settings are the same, and the ε value is set to 0.8.

[0117] Table 1

[0118]

[0119] First, regarding the estimation of the deployment location of the connector, this experiment sets the population size to 30 and the maximum number of iterations to 500. Generally speaking, in natural disasters or emergencies, the area where communication is restored reaches a certain proportion, which means that the communication is restored successfully. This proportion can be determined based on the geographical scope, population size, or the requirements of specific key places. The actual scale threshold may vary depending on the type of disaster and the emergency situation. Regarding the communication recovery connection problem in this article, there is theoretically no limit on the number of wireless network connectors, and the communication reconnection rate can be as high as 100%. However, due to the limitations of wireless network connectors, a standard is set that the communication reconnection rate reaches 90% to be considered a successful connection. Based on the above, the closer the communication reconnection rate is to 100%, the better, provided that the requirements are met.

[0120] The results of 25 independent runs are shown in Table 2. Table 2 lists the connectivity rates obtained by different algorithms in the experiment, which can more clearly reflect the algorithm's optimization ability for the location of wireless network connectors. We use the average connectivity rate and the optimal connectivity rate to measure the algorithm results. The Best index represents the best result of the connectivity rate obtained by using a certain intelligent optimization algorithm in these 25 experiments, reflecting the optimization ability of the algorithm. The Mean index represents the average value of the connectivity rate results obtained by using a certain intelligent optimization algorithm in these 25 experiments, reflecting the average level of algorithm optimization. From the experimental results, it can be seen that among the results obtained by the five algorithms of PSO, HHO, SMA, MPA, and ε-MPA, the algorithm obtained by the ε-MPA algorithm is the best, which is 0.77% higher than the second-ranked MPA algorithm. Since these five algorithms are all meta-heuristic algorithms, which are used to obtain satisfactory solutions within an acceptable time, they are approximate solution algorithms with a certain randomness, so not every run can reach the global optimal solution. It can be seen that the average index and connectivity rate obtained by the ε-MPA algorithm are the best, and the optimal connectivity rate is 97.89%. Therefore, the ε-MPA algorithm is used to estimate the position of the wireless connector. When the coverage rate reaches the optimal value in the 14th operation, its position is as follows: Figure 6 shown.

[0121] Table 2: Reconnaissance coverage results of five optimization algorithms

[0122]

[0123] Next, for task clustering, the number of drones was set in the experiment, and the K-means algorithm was used for clustering. The number of clusters was set to be the same as the number of drones, and the initial cluster centers were randomly selected. The clustering results are shown in the figure below. Figure 7 As shown. In order to verify the impact of different numbers of drones on clustering results, we set the number of clusters to 1 to 4. For the K-means algorithm, the selection of the initial cluster center will also affect the experimental results, so we choose a random method to eliminate errors. According to the results of different numbers of clusters, it can be seen that the clustering results are good and the distribution is reasonable. The clustering results will determine the location and number of network connectors that each drone needs to transport, which is also a necessary condition for drone path planning.

[0124] Finally, for the UAV path planning, when obtaining the UAV task clustering results, the swarm intelligence algorithm is continued to be used for motion planning, with a maximum number of iterations of 500 times. The results of 25 independent runs are shown in Table 3. HHO, SMA, MPA and ε-MPA can all achieve the same optimal level. EMPA has the strongest optimization ability in most scenarios, MPA has the best average optimization effect, and ε-MPA is second. Since the greedy strategy is used to solve the exploration and utilization problem, over-emphasizing exploration will cause certain fluctuations.

[0125] Table 3: Total distance traveled by the drone swarm

[0126]

[0127] In scenario 1, when there are 4 drones and the optimal planning is achieved, starting from different starting points, the paths for each drone to complete the task are 677.153, 672.089, 844.839, and 946.268 respectively. The paths of each drone are as follows: Figure 8 As shown on the left. The total length of the path used by the drone group to complete the network connection task is 3140.35km. In scenario 2, when the number of drones is 4 and the optimal plan is obtained, starting from the same starting point (250, 0), the paths of each drone to complete the task obtained by the EMPA algorithm are 1302.005, 864.092, 929.326, and 1318.693 respectively. The paths of each drone are shown in Figure 8 As shown on the right, the total mileage of the drone swarm to complete the task of deploying wireless network connectors is 4414.116 kilometers.

[0128] In the above-mentioned multi-UAV network communication task planning method, a multi-UAV network communication task planning model is established for the background of network connectivity problems in harsh environments such as earthquakes and fires, and a K-means clustering combined with the ε-MPA algorithm is proposed. Through wireless network connector position estimation, multi-UAV task clustering and motion planning, efficient network communication in certain disaster environments is achieved. The wireless connector position estimation and delivery path planning results of PSO, HHO, SMA, MPA and ε-MPA are compared in the experiment, indicating that ε-MPA has the best performance in most cases and can maximize the network connectivity range under the shortest flight distance of the UAV group. This method uses the ε-greedy strategy to take into account exploration and utilization in the optimization process, further improving the optimization ability of MPA. The ε-MPA algorithm has achieved the best performance in solving the problem of wireless connector position estimation and has achieved the best performance in multiple scenarios of multi-UAV path planning. As a meta-heuristic algorithm for solving approximate solutions, ε-MPA is suitable for various problems with complex search spaces and optimization objectives such as scheduling, organizational optimization, and resource allocation. It can help find optimal or suboptimal solutions and optimize decision-making and resource utilization.

[0129] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0130] In one embodiment, Fig. 9 As shown, a network communication task planning device based on multiple UAVs is provided, including: a communication task information acquisition module 200, a delivery task location acquisition module 210, a UAV task allocation module 220 and a UAV task route planning module 230, wherein:

[0131] The communication task information acquisition module 200 is used to obtain the area of ​​the area to be restored, the number of connector nodes, and the initial positions of multiple drones;

[0132] The delivery task location obtaining module 210 is used to find a set of connector node setting locations that maximize the connectivity area in the connectivity area to be restored, and use the set of locations as the target locations for the multi-UAV delivery task;

[0133] The UAV task allocation module 220 is used to cluster the multiple UAV delivery tasks with the shortest total flight path of each UAV as the goal, using the initial position of each UAV as the cluster center, and obtain a delivery task set corresponding to each UAV;

[0134] The UAV task route planning module 230 is used to plan the path for each UAV to complete all delivery tasks in the corresponding delivery task set using the optimized marine predator algorithm for each UAV to obtain the network communication task plan for each UAV.

[0135] For the specific definition of the network communication task planning device based on multiple drones, please refer to the definition of the network communication task planning method based on multiple drones above, which will not be repeated here. Each module in the above-mentioned network communication task planning device based on multiple drones can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0136] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig.10As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a network communication task planning method based on multiple unmanned aerial vehicles is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0137] Those skilled in the art will understand that Fig.10 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0138] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0139] Obtain the area of ​​the connection area to be restored, the number of connector nodes, and the initial positions of multiple drones;

[0140] Find a group of connector node settings in the area to be restored that maximizes the connectivity area, and use the group of locations as the target locations for the multi-UAV delivery mission;

[0141] With the goal of minimizing the total flight path of each drone, the initial position of each drone is used as a clustering center to cluster the multiple drone delivery tasks to obtain a delivery task set corresponding to each drone;

[0142] For each of the delivery task sets, the optimized ocean predator algorithm is used to plan the paths for each drone to complete all delivery tasks in the corresponding delivery task set, so as to obtain the network communication task plan for each drone.

[0143] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0144] Obtain the area of ​​the connection area to be restored, the number of connector nodes, and the initial positions of multiple drones;

[0145] Find a group of connector node settings in the area to be restored that maximizes the connectivity area, and use the group of locations as the target locations for the multi-UAV delivery mission;

[0146] With the goal of minimizing the total flight path of each drone, the initial position of each drone is used as a clustering center to cluster the multiple drone delivery tasks to obtain a delivery task set corresponding to each drone;

[0147] For each of the delivery task sets, the optimized ocean predator algorithm is used to plan the paths for each drone to complete all delivery tasks in the corresponding delivery task set, so as to obtain the network communication task plan for each drone.

[0148] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0149] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0150] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A network communication task planning method based on multiple UAVs, characterized in that: The method comprises: Obtain the area of ​​the connection area to be restored, the number of connector nodes, and the initial positions of multiple drones; Find a group of connector node settings in the area to be restored that maximizes the connectivity area, and use the group of locations as the target locations for the multi-UAV delivery mission; With the goal of minimizing the total flight path of each drone, the initial position of each drone is used as a clustering center to cluster the multiple drone delivery tasks to obtain a delivery task set corresponding to each drone; For each of the delivery task sets, the optimized ocean predator algorithm is used to plan the paths for each drone to complete all delivery tasks in the corresponding delivery task set, so as to obtain the network communication task plan for each drone.

2. The network communication task planning method according to claim 1, characterized in that: The step of finding a set of connector node locations that maximize the connectivity area in the connectivity area to be restored comprises: Divide the area to be restored into a plurality of circular sub-areas with different areas, each of which is represented on a two-dimensional plane as the coordinates of a center point and a radius; Obtaining the connected area of ​​each of the connector nodes according to the position coordinates and the sensing radius of each of the connector nodes; Determine whether the sub-area is covered by the connector node according to the distance from the center of the sub-area to the center of the connector node and the connector sensing radius; The regional connectivity rate of all connector nodes in the connectivity area to be restored is defined as the ratio of the unconnected sub-area covered by the connector node set to the total area of ​​the connectivity area to be restored; Find a set of connector node coordinates that maximizes the connectivity rate of the area.

3. The network communication task planning method according to claim 2, characterized in that: A 0-1 perception model is used to obtain a set of connector node coordinates that maximizes the connectivity rate of the area.

4. The network communication task planning method according to claim 3, characterized in that: When clustering delivery tasks, the k-means clustering algorithm is used.

5. The network communication task planning method according to any one of claims 1 to 4, characterized in that: When the optimized marine predator algorithm is used for path planning for each UAV, the objective function is expressed as: In the above formula, the set of delivery tasks corresponding to a certain drone is expressed as V = {v1, v2, v3, ..., v A }, the order of placing the connector nodes is expressed as T = {t1, t2, t3, ..., t A }, where t i =V(i=1,…,A), and t i+1 =t1.

6. The network communication task planning method according to claim 5, characterized in that: The optimized marine predator algorithm combines the ε-greedy strategy on the basis of the marine predator algorithm, and determines the adopted strategy according to the preset ε value and the random number in different iteration stages.

7. The network communication task planning method according to claim 6, characterized in that: The optimized marine predator algorithm includes: In the initial stage of iteration, if the random number is less than the preset ε value, the first strategy is used to search in a large range. If the random number is greater than or equal to the preset ε value, the second strategy or the third strategy is randomly used to search. In the middle stage of the iteration, if the random number is less than the preset ε value, the second strategy is randomly adopted to search in a small range; if the random number is greater than or equal to the preset ε value, the first strategy or the third strategy is randomly adopted to search; At the end of the iteration, if the random number is less than the preset ε value, the third strategy is randomly adopted to search in the local range; if the random number is greater than or equal to the preset ε value, the first strategy or the second strategy is randomly adopted to search.

8. A network communication task planning device based on multiple UAVs, characterized in that: The device comprises: The communication task information acquisition module is used to obtain the area of ​​the connection area to be restored, the number of connector nodes, and the initial positions of multiple drones; A delivery task location obtaining module is used to find a set of connector node setting locations that maximize the connectivity area in the connectivity area to be restored, and use the set of locations as the target locations for the multi-UAV delivery task; The UAV task allocation module is used to cluster the multiple UAV delivery tasks with the shortest total flight path of each UAV as the goal, using the initial position of each UAV as the clustering center, and obtaining a delivery task set corresponding to each UAV; The UAV task route planning module is used to plan the path for each UAV to complete all delivery tasks in the corresponding delivery task set using the optimized marine predator algorithm for each UAV to obtain the network communication task plan for each UAV.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images