Unmanned aerial vehicle cluster formation configuration and spectrum resource joint optimization method, system, device, medium and product

Abstract

The invention discloses an unmanned aerial vehicle cluster formation configuration and spectrum resource joint optimization method, system, device, medium and product, and relates to the field of unmanned aerial vehicle planning control, and the method comprises the steps: dispatching a plurality of unmanned aerial vehicle formations to execute a plurality of tasks under the conditions of given spectrum resources and limited number of unmanned aerial vehicles, constructing an optimization problem model by using a coupling relationship among the number of the unmanned aerial vehicle formations, a target sensing sequence and spectrum resource allocation; solving the optimization problem model by using an improved genetic algorithm, and outputting an optimal unmanned aerial vehicle cluster formation configuration and spectrum resource joint optimization scheme; according to the improved genetic algorithm, a simulated annealing algorithm is embedded into the genetic algorithm, a crossover operator of the genetic algorithm is adjusted, and an elite strategy and a dynamically adjustable mutation rate are introduced into mutation operation. According to the invention, joint planning of unmanned aerial vehicle task planning, unmanned aerial vehicle formation number distribution and bandwidth selection can be realized, and the method has higher global optimization capability.

Description

Technical Field

[0001] The present application relates to the field of unmanned aerial vehicle (UAV) planning and control, and in particular to a method, system, device, medium and product for joint optimization of UAV cluster formation configuration and spectrum resources. Background Art

[0002] Unmanned aerial vehicles (UAVs) are widely used in battlefield reconnaissance, joint attacks, emergency rescue and other operations due to their low cost, high flexibility and strong concealment. Compared with single drones, drone clusters have more advantages in mission execution, can perform more diverse tasks, and conduct large-scale searches more efficiently. Multi-UAV collaborative mission execution is aimed at two or more drones to collaboratively perform tasks in complex environments. Compared with single drones, multi-UAV collaborative mission execution has better flexibility in handling tasks and is more suitable for performing complex tasks. Although multi-task systems have great advantages, multi-UAVs face problems such as how to balance the relationship between mission requirements and resource planning. If only focusing on mission requirements, it may lead to a waste of UAV resources and spectrum resources; while excessive focus on resource allocation will lead to insufficient mission completion, resulting in reduced benefits; it can be seen that the mission requirements and resource allocation of drones are coupled with each other, and the two jointly determine the optimal drone collaborative mission planning scheme.

[0003] Existing research often only focuses on meeting the needs of the task and ignores the resource level. In terms of meeting the needs of the task, domestic and foreign scholars have done a lot of research. For example, 1) By establishing the overall efficiency model of task allocation under the factors of the benefits of executing the task, the time of executing the task, and the task load, a quantum genetic algorithm based on the idea of ​​distribution estimation is proposed, which realizes the initial allocation of drone tasks in multi-base situations and dynamic allocation under sudden threats. 2) For the problem of multi-drone task allocation in multiple target tracking scenarios, a distributed allocation method based on the auction pigeon swarm optimization (auction-PIO) algorithm is proposed. The proposed algorithm has good task allocation performance, and outdoor experiments verify that the algorithm can meet the practical requirements. 3) In the face of the problem of dynamic task allocation of drone swarms in complex scenarios, the behavioral characteristics of wolf packs in natural environments are modeled and solved using this feature. A path planning algorithm and coverage search algorithm suitable for drone swarms in complex scenarios are also proposed, which achieves a high degree of task completion, balances the load of drones, and provides a suitable solution for complex scenarios. 4) In the face of the coverage path planning problem of heterogeneous UAVs, a heuristic method based on the ant colony algorithm (ACS) is proposed to seek approximate optimal solutions and minimize the time consumption of tasks in collaborative search systems. 5) In the face of the collaborative search and attack mission planning problem of multi-UAVs, an intelligent self-organizing algorithm (ISOA) is proposed. The proposed algorithm has stable and efficient convergence performance, external responsiveness and internal scalability. 6) In the face of the problems of limited UAV flight time and limited load capacity in the mission planning process, the E-constraint method is used to derive the exact Pareto optimal solution. In order to find an approximate Pareto solution, an approximate two-phase (ATP) method is developed, which has the characteristics of high accuracy and shortens the calculation time by two orders of magnitude.

[0004] Reasonable spectrum resource optimization can effectively reduce the energy loss of drones. Existing research has also done a lot of work on the optimization of spectrum resources. 1) In the face of the problem of drone-assisted user communication, the problem is modeled as a mixed integer non-convex optimization problem, which is decomposed into integer optimization sub-problems and non-convex optimization sub-problems according to the different characteristics of discrete and continuous optimization variables. Then an iterative algorithm combining genetic algorithm and convex optimization is developed to alternately optimize these two separate sub-problems. The algorithm jointly optimizes user association and drone position to maximize the total rate of downlink users. The results show that the algorithm has good robustness. 2) In the face of the problem of limited carrying capacity of a single drone for wireless communication of unmanned aerial vehicles (UAVs). Multiple drones are dispatched to collect data from multiple ground nodes, and the energy consumption of drones is minimized by jointly optimizing the trajectory and node communication time. 3) In the face of the scenario of drone clusters as auxiliary deployment of wireless networks, by constructing a mean potential field game, the model is solved using the primal double hybrid gradient (PDHG) method, while minimizing energy consumption, high channel capacity is achieved.

[0005] The above existing technologies often only focus on mission requirements or spectrum resource planning, and there is little research on the combination of the two. Especially as the scale of tasks to be completed becomes larger and larger, the complexity of tasks continues to increase, more and more drones are used, and the spectrum becomes more and more crowded, unreasonable resource allocation will reduce the quality of mission completion. Summary of the invention

[0006] The purpose of this application is to provide a method, system, device, medium and product for joint optimization of drone cluster formation configuration and spectrum resources to solve the problem of not being able to take into account both drone formation and spectrum resources.

[0007] To achieve the above objectives, this application provides the following solutions:

[0008] In a first aspect, the present application provides a method for jointly optimizing UAV swarm formation configuration and spectrum resources, including:

[0009] Under the conditions of given spectrum resources and limited number of drones, multiple drone formations are dispatched to perform multiple tasks. The optimization problem model is constructed by using the coupling relationship between the number of drone formations, target perception order and spectrum resource allocation; wherein the target perception order is the execution order of each drone formation to perform tasks; one target is executed by one drone formation;

[0010] The optimization problem model is solved by an improved genetic algorithm, and an optimal UAV cluster formation configuration and spectrum resource joint optimization scheme is output; wherein, the improved genetic algorithm is to embed a simulated annealing algorithm in the genetic algorithm, adjust the crossover operator of the genetic algorithm, and introduce an elite strategy and a dynamically adjustable mutation rate in the mutation operation; the crossover operator of the genetic algorithm is adjusted to inherit the characteristics of different parent generations while ensuring the correctness of the offspring; the elite strategy is used to retain the best individuals and directly enter the next generation without crossover and mutation operations; the dynamically adjustable mutation rate is used to dynamically adjust the mutation rate according to the operation of the improved genetic algorithm, and the mutation rate is used to explore the solution space and adjust the solution; the simulated annealing algorithm is used to accept poor solutions in the high temperature stage and gradually converge to the optimal solution from the poor solutions in the low temperature stage; the optimal solution is the optimal UAV cluster formation configuration and spectrum resource joint optimization scheme; the poor solution is a solution with worse quality than the current solution.

[0011] In the second aspect, the present application provides a system for configuring a drone cluster formation and jointly optimizing spectrum resources, including:

[0012] The optimization problem model building module is used to dispatch multiple drone formations to perform multiple tasks under the conditions of given spectrum resources and a limited number of drones, and to build an optimization problem model by using the coupling relationship between the number of drone formations, the target perception order, and the spectrum resource allocation; wherein the target perception order is the execution order of each drone formation to perform tasks; and one target is executed by one drone formation;

[0013] The optimal UAV swarm formation configuration and spectrum resource joint optimization solution output module is used to solve the optimization problem model by using an improved genetic algorithm, and output the optimal UAV swarm formation configuration and spectrum resource joint optimization solution; wherein, the improved genetic algorithm is to embed a simulated annealing algorithm in the genetic algorithm, adjust the crossover operator of the genetic algorithm, and introduce an elite strategy and a dynamically adjustable mutation rate in the mutation operation; adjusting the crossover operator of the genetic algorithm can inherit the characteristics of different parent generations while ensuring the correctness of the offspring; the elite strategy is used to retain the best individuals and directly enter the next generation without crossover and mutation operations; the dynamically adjustable mutation rate is used to dynamically adjust the mutation rate according to the operation of the improved genetic algorithm, and the mutation rate is used to explore the solution space and adjust the solution; the simulated annealing algorithm is used to accept poor solutions in the high temperature stage, and gradually converge from the poor solutions to the optimal solution in the low temperature stage; the optimal solution is the optimal UAV swarm formation configuration and spectrum resource joint optimization solution.

[0014] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the above-mentioned methods for joint optimization of drone cluster formation configuration and spectrum resources.

[0015] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the above-mentioned methods for unmanned aerial vehicle cluster formation configuration and spectrum resource joint optimization.

[0016] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of any of the above-mentioned methods for drone cluster formation configuration and spectrum resource joint optimization.

[0017] In order to rationally utilize UAV resources and maximize the benefits of mission completion, taking into account the limited mileage that UAVs can fly, this application jointly optimizes the "reconnaissance-communication" mission planning and spectrum resources of multiple UAVs in rescue operations in a multi-task scenario with limited resources, and constructs an optimization problem model by utilizing the coupling relationship between the number of UAV formations, target perception order, and spectrum resource allocation; an improved genetic algorithm is designed to solve the optimization problem model to obtain the optimal UAV cluster formation configuration and spectrum resource joint optimization solution. Among them, the improved genetic algorithm is to embed a simulated annealing algorithm in the genetic algorithm, adjust the crossover operator of the genetic algorithm, and introduce an elite strategy and a dynamically adjustable mutation rate in the mutation operation; the elite strategy is used to retain the best individuals and directly enter the next generation without undergoing crossover and mutation operations; the dynamically adjustable mutation rate is used to dynamically adjust the mutation rate according to the operation of the improved genetic algorithm, and the mutation rate is used to explore the solution space and adjust the solution; the simulated annealing algorithm is used to accept poor solutions in the high temperature stage and gradually converge to the optimal solution from the poor solution in the low temperature stage; the optimal solution is the optimal UAV cluster formation configuration and spectrum resource joint optimization scheme, the improved algorithm combines the large-scale exploration capability of the genetic algorithm and the fine search advantage of the simulated annealing algorithm, realizes the joint planning of UAV task planning, UAV formation number allocation and bandwidth selection, and has a higher global optimization capability. Compared with other heuristic algorithms, the task selection and UAV formation number planned by the improved genetic algorithm are more reasonable, and can cope with more complex scene environments, and achieve greater network benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 This is a flow chart of a method for joint optimization of drone cluster formation configuration and spectrum resources in one embodiment of the present application;

[0020] Figure 2 A schematic diagram of a scenario in which a drone performs an integrated “reconnaissance-communication” mission provided by an embodiment of the present application;

[0021] Figure 3 for Figure 1 A detailed flow chart of step 102;

[0022] Figure 4 A schematic diagram of a three-chromosome scheme for a formation of three drones to execute 10 targets provided in an embodiment of the present application;

[0023] Figure 5 A schematic diagram of a crossover operation provided in an embodiment of the present application;

[0024] Figure 6 A schematic diagram of the mileage limit of a drone formation provided in an embodiment of the present application;

[0025] Figure 7 A detailed flowchart of step 102 provided in an embodiment of the present application;

[0026] Figure 8 A drone quantity planning diagram for drone mission planning of 10 drones with a mission number of 8 assigned to three airport drone formations provided in an embodiment of the present application;

[0027] Fig. 9 A drone quantity planning diagram for drone mission planning in which 10 drones with a mission number of 10 are assigned to three airport drone formations in an embodiment of the present application;

[0028] Fig.10 A frequency selection diagram for drone mission planning of 10 drones with a mission number of 8 assigned to 3 airport drone formations provided in an embodiment of the present application;

[0029] Fig.11 A frequency selection diagram for drone mission planning of 10 drones with a mission number of 10 assigned to 3 airport drone formations provided in an embodiment of the present application;

[0030] Fig.12 The result graph of different algorithms for the task number of 8 and the allocation of 10 drones to the drone formation of 3 take-off points provided for an embodiment of the present application;

[0031] Fig.13 The number of tasks provided for an embodiment of the present application is 10, and a total of 10 drones are assigned to a drone formation at 3 take-off points. Result diagram of different algorithms;

[0032] Fig.14 A graph showing the revenue results for different numbers of drones under two different weights provided in one embodiment of the present application;

[0033] Fig.15 A comparison chart of algorithm stability under the same resources provided in an embodiment of the present application when a take-off point is added. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0035] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0036] like Figure 1 As shown, the embodiment of the present application provides a method for jointly optimizing UAV cluster formation configuration and spectrum resources, including:

[0037] Step 101: Under the conditions of given spectrum resources and a limited number of drones, multiple drone formations are dispatched to perform multiple tasks, and an optimization problem model is constructed using the coupling relationship between the number of drone formations, target perception order, and spectrum resource allocation; wherein the target perception order is the execution order of each drone formation performing a task; and one target is executed by one drone formation.

[0038] Step 102: Solve the optimization problem model using an improved genetic algorithm, and output an optimal UAV cluster formation configuration and spectrum resource joint optimization solution; wherein the improved genetic algorithm is to embed a simulated annealing algorithm in the genetic algorithm, adjust the crossover operator of the genetic algorithm, and introduce an elite strategy and a dynamically adjustable mutation rate in the mutation operation; adjusting the crossover operator of the genetic algorithm can inherit the characteristics of different parent generations while ensuring the correctness of the offspring, avoiding the omission and duplication of the target; the elite strategy is used to retain the best individuals and directly enter the next generation without undergoing crossover and mutation operations; the dynamically adjustable mutation rate is used to dynamically adjust the mutation rate according to the operation of the improved genetic algorithm, and the mutation rate is used to explore the solution space and adjust the solution; the simulated annealing algorithm is used to accept poor solutions in the high temperature stage, and gradually converge from the poor solutions to the optimal solution in the low temperature stage; the optimal solution is the optimal UAV cluster formation configuration and spectrum resource joint optimization solution.

[0039] In an exemplary embodiment, Figure 2 As shown in the figure, the scenario of drones performing the integrated "reconnaissance-communication" task is described. Specifically, there are multiple heterogeneous drones in the scenario, and the drones carry reconnaissance and communication task payloads. As shown in the figure, consider the scenario of drone clusters performing "reconnaissance and communication" tasks. There are multiple different airports as take-off points, and there are multiple "reconnaissance-communication" tasks to be performed. The drones in the same airport form a drone formation, which traverses and performs the assigned tasks. After completing the reconnaissance task of a fixed time, they hover above the task point and transmit the reconnaissance information to the airport's decision center for processing. Since the battery capacity of drones is relatively small and cannot support long-term working energy consumption, when there are many tasks, a single drone formation may not be able to complete all "reconnaissance-communication" tasks. Set up multiple nearby airports and drone formations. Assume that M drones need to be deployed to N different airports, and each airport dispatches drones to form a drone formation, and the airport is the take-off point to go to K task point targets distributed on the ground to perform cluster reconnaissance-communication tasks. The sets of drones, airports, and targets are M0 = {1, ..., M}, N0 = {1, ..., N}, and K0 = {1, ..., K}, respectively.

[0040] Assume that the total system bandwidth B is divided into N sub-channels, which are managed by the base station. The drones in the alliance use the channels in turn to transmit information to the local departments. Different alliances use different channels. In the same alliance, each drone uses the same channel through time division multiplexing. Due to the huge differences in weather and topography in the disaster-stricken areas, the difficulty of drone reconnaissance missions varies under different tasks. The relevant departments have connected with the local meteorological bureau in advance and evaluated the weather and topography. The completion probability of a drone when performing task i can be set to P. i . Let the tuple (Val(i), λ i 、T A (i))) is the attribute of task i, where (Val(i), λ i 、T A (i)) denote the value of task i, the amount of data transferred by task i and the fixed completion time respectively.

[0041] Drone Cluster A i The probability of completing task i is:

[0042]

[0043] Assuming that the value of a task increases as the probability of task completion increases, the value of task i can be expressed as:

[0044] V(i)=P(A i ,i)*Val(i)

[0045] The horizontal coordinate of the ground-based target is q k =R 2*1 Represents and is known. The distance between target k and the decision center can be expressed as:

[0046] d k,n =||q k -q n ||

[0047] Among them, q n represents the horizontal coordinate of the nth airport decision center, q k represents the horizontal coordinate of the kth task point, and ||*|| represents the Euclidean distance.

[0048] Assume that a small drone is used to perform a "reconnaissance-communication" mission. It can hover in the air to obtain stable communication transmission quality. The energy consumption of a drone is divided into two parts: motion energy consumption and communication energy consumption. Generally speaking, the communication energy consumption is several orders of magnitude lower than the motion energy consumption and can be ignored. The motion energy consumption of the drone is mainly considered. When the drone moves at a constant speed of v, its power consumption can be expressed as:

[0049]

[0050] p0 and p i They represent the blade profile power and induced power in the hovering state respectively; U tip and v0 represent the rotor tip speed and the rotor average speed in the hovering state respectively; d0 and s represent the fuselage drag ratio and the rotor solidity respectively; ρ and A represent the air density and the rotor disc area respectively.

[0051] When the distance between targets is far, the acceleration and deceleration time of the UAV is very short compared to the time of maneuvering between targets. The acceleration and deceleration process of the UAV can be ignored. The UAV adopts a uniform flight speed when maneuvering between targets and hovers when performing perception tasks.

[0052] The energy consumption of drones at different flight speeds is different. It can be verified from the formula that the flight power consumption of drones first decreases and then increases with the increase of speed. There is a maximum cruising range speed V mr .

[0053] To save energy, it is assumed that the UAV uses a speed V when flying between targets. mr , so the minimum flight energy consumption required for the drone to fly to the next target k' after completing the perception task of target k is:

[0054]

[0055] represents the time it takes for the drone to fly from target k to k'; d k.k' It represents the distance between target k and k', which is calculated by the formula.

[0056] Due to the limited battery power of the drone, the hovering time is calculated based on the information return time of the drone. Thanks to the drone's lift gain, the air-to-ground channel power gain is mainly determined by the line-of-sight link and can be expressed based on the free space path loss model as:

[0057]

[0058] Where β0 represents the channel power gain at a reference distance of 1m. k,c It is the distance from the UAV formation to the starting point.

[0059] Drone formation N i The bandwidth allocated is W i , the fixed amount of information transmitted by task k is C k The hovering time of the drone at target k is:

[0060]

[0061] σ 2 Represents the ambient noise power, Pm Represents the transmission power of UAV m.

[0062] The energy consumption of the drone hovering at target k is:

[0063]

[0064] Formula, P(0) is the flight energy consumption when the UAV speed is 0, that is, the flight power energy consumption when hovering. i is the fixed completion time of task i.

[0065] One target only needs one UAV formation to perform the task. The association between target k and UAV formation n is:

[0066]

[0067] The total distance S of any UAV formation n n for:

[0068]

[0069] Total distance

[0070] in, Indicates the distance from the airport of UAV formation i to the first selected mission point, Indicates the last mission point to team n's airport π n (0) distance; K n is the total number of tasks assigned to UAV formation n; represents the kth mission point π selected by the UAV formation n n (k) to the k+1th task point π n (k+1) distance; |A n | represents the number of drones in the drone formation n; N is the total number of drones in the drone formation.

[0071] The optimization goal of the present invention is to dispatch drone formations to perform information "reconnaissance and communication" tasks under the conditions of given spectrum resources and limited number of drones, and to deploy the number of drones for drone formations with different starting points, allocate bandwidth, reasonably allocate tasks and plan task paths. Multiple drones collaborate to complete reconnaissance and communication tasks, which is essentially a constrained multiple traveling salesmen problem (MTSP). Different drone formations start from different starting points, and each task only requires one drone formation to perform.

[0072] A single drone formation cannot complete large-scale reconnaissance and communication tasks, and requires the collaboration of multiple drone formations. In actual scenarios, most people only consider the degree of task completion, task completion time, or drone demand or resource loss, and choose one aspect for consideration. In reality, the relationship between benefits and energy consumption should be reasonably balanced. Use weight coefficients to combine them. Reasonable drone formation execution order and spectrum allocation can improve mission efficiency, reduce drone spectrum waste, and thus reduce drone energy consumption. Therefore, the optimization goal can essentially be transformed into optimizing the number of drones, the execution order of each drone formation, and spectrum resource allocation under the condition of a given number of drone formations, so as to maximize the total utility function of multiple tasks. The mathematical model of the optimization problem is expressed as:

[0073]

[0074] C4: S≤S max

[0075] C5:|A n |>0

[0076]

[0077] Among them, U is the total utility function; k n is the task number assigned to formation n; N is the number of formations; p k is the completion probability of a drone completing task k; |A n | is the number of drones in formation n; V(k) is the value of the task; γ is the weight coefficient of the hovering loss; P(v) is the power loss function when the drone speed is v. When the drone is hovering, the speed is 0, then P(0) is the unmanned power loss during hovering; T(k) is the fixed reconnaissance time of task k, λ k is the amount of information data that needs to be transmitted for task k; W n is the bandwidth of formation n; P m is the transmitting power of the UAV; is the distance from the mission point of UAV formation n to the airport of formation n when UAV formation n performs mission k; β0 is the channel power gain at a distance of 1m, σ 2 is the power spectrum density of environmental noise; λ is the weight coefficient of path loss; V mr The flight speed at which the drone cruises the longest distance; is the time for the UAV formation n to traverse the mission point; X k,n Perform task k for UAV formation n; n is the sequence number of the UAV formation; K is the total number of tasks; M is the number of UAVs that can be assigned; S max is the maximum mileage of the drone; X n is the formation number for the mission; M is the total number of drones that can be assigned; Wsum is the total bandwidth of the drone formation; C1 indicates that each task corresponds to a formation, C2 all tasks are performed by the drone formation, C3 the sum of the drones in each formation is the number to be allocated, C4 indicates that the formation mileage must be less than the limit value, C5 indicates that each formation must be allocated drones, C6 indicates that the sum of the bandwidths of each formation is the existing bandwidth, and C7 indicates that the bandwidth of each formation is greater than 0. This is a mixed-integer nonlinear programming (MINLP) problem. The variables to be optimized are coupled with each other and are difficult to solve using existing optimization methods. Solving the optimal solution is challenging. In an exemplary embodiment, genetic algorithms usually have a huge search space for optimization problems of discrete variables. Through random generation and evolution, they can effectively search a large-scale candidate solution space to find potential optimal solutions; a simulated annealing algorithm is introduced. The role of simulated annealing in the genetic algorithm is to further optimize individuals to help escape from local optimal solutions. Figure 3 As shown, step 102 can be replaced by the following steps 201-205:

[0078] Step 201: Use a chromosome to initialize the execution task plan of the drone formation to generate the first generation of individuals; wherein a chromosome includes four segments, the first segment of the chromosome with a length of K represents the perception order of each target, the second segment of the chromosome with a length of K represents the tasks assigned to each airport, the third segment of the chromosome with a length of N represents the number of drones in each drone formation, and the fourth segment of the chromosome with a length of N represents the bandwidth allocated to each drone formation.

[0079] Furthermore, Figure 4 It shows three chromosome schemes for three drone formations to execute 10 targets, where drone formation 1 senses 4 targets in the order of 1, 4, 5 and 2, the number of drones in drone formation 1 is 3, and the bandwidth is divided into standard one, and the same is true for other drone formations.

[0080] Step 202: Introduce an elite strategy to select the best individuals from the first generation of individuals and directly use them as the next generation of individuals without undergoing crossover and mutation operations.

[0081] Step 203: The remaining individuals in the first generation are subjected to crossover operation, mutation operation and drone mileage limit in turn to generate the next generation of individuals, and the optimal individuals are combined to generate a new generation of individuals; wherein the drone mileage limit is used to perform feasibility processing on the new schemes generated after the crossover operation and mutation operation, and to determine the mileage of each drone formation to perform the task.

[0082] Furthermore, each operation in the improved genetic algorithm is explained.

[0083] Crossover operation: Crossover operation creates new offspring individuals by mixing the genetic information of parent individuals. It encourages the genetic algorithm to explore diversity in the solution space. The task target number is unique and directional. Traditional crossover operations are prone to duplication and omission of targets. In order to prevent this problem, this application adjusts the crossover operator. While inheriting the characteristics of different parent generations, it also ensures the correctness of the offspring, avoids the omission and duplication of targets, and avoids the omission and duplication of targets. Specific operations are as follows: Figure 5 shown.

[0084] Mutation operation: Mutation operation is a random operation that introduces new gene combinations by randomly changing individual genes to increase the diversity of the population. This application focuses on the allocation of the number of drone formations that perform tasks collaboratively, and needs to consider the number of drones in the drone formation, the order of performing tasks, the number of drone tasks performed, and the bandwidth allocation of each drone formation.

[0085] Mileage limit of drone formation: In order to meet the energy consumption requirements of drones, it is necessary to conduct feasibility processing on the new solutions generated by cross-mutation, and it is necessary to determine the mileage of each drone formation to perform tasks. Set a penalty value. If the mileage value of the drone formation's traversal task exceeds the set value, the penalty value is subtracted from the fitness function. Then, a task in the drone formation is assigned to other drone formations, and the mileage value is re-determined. Figure 6 shown.

[0086] Elite strategy and dynamically adjustable mutation rate: In this optimization problem, there are many variables to be optimized and the permutations and combinations of variables are huge. In order to improve the solving performance of the genetic algorithm, the elite strategy and dynamically adjustable mutation rate are introduced in the selection and mutation operations. The elite strategy refers to retaining a part of the best-performing individuals (i.e., elites) in the selection process of the genetic algorithm and directly entering the next generation without crossover and mutation operations. The mutation rate is a mechanism used to introduce new individuals in the genetic algorithm, which determines the probability of individual genes mutating in each generation. The dynamically adjustable mutation rate strategy refers to dynamically adjusting the size of the mutation rate according to the operation of the algorithm. The purpose of the dynamically adjustable mutation rate is usually to balance exploration and exploitation. In the early stage of the algorithm, a higher mutation rate helps to explore the solution space; in the later stage of the algorithm, a lower mutation rate helps to fine-tune the solution to avoid destroying the good solution that has been found. They ensure that the algorithm can effectively converge to the optimal solution while maintaining its exploration ability.

[0087] Simulated annealing algorithm: As the number of iterations increases, the differences between each individual in the population also decrease, which leads to a decrease in the convergence of the genetic algorithm and gradually falls into the local optimal solution. To avoid this situation, by embedding simulated annealing in the genetic algorithm, individuals in each generation can be locally optimized. Simulated annealing accepts poor solutions in the high temperature stage and gradually converges to a better solution in the low temperature stage, which helps the genetic algorithm jump out of the local optimal solution and improves the global search capability.

[0088] Step 204: taking the new generation of individuals as the initial solution of the simulated annealing algorithm, and evaluating the fitness of the initial solution.

[0089] Step 205: Based on the fitness of the initial solution, the new generation of individuals is iteratively trained until the maximum number of iterations is reached, and the optimal drone cluster formation configuration and spectrum resource joint optimization solution is output.

[0090] In another exemplary embodiment, step 102 may be replaced by the following steps 301 to 307:

[0091] Step 301: Initialize the population and divide the variables to be solved into four chromosomes to form a genetic algorithm population, which are the execution order of each target, the tasks assigned to each drone formation, the number of drones in each drone formation, and the bandwidth selection of each drone formation.

[0092] Step 302: The elite selection operation retains a portion of the best performing individuals (i.e., elites) to directly enter the next generation without undergoing crossover and mutation operations.

[0093] Step 303: Crossover operation, randomly switching the arrangement of the first chromosome, i.e. the execution order of the drones, to generate more types of chromosomes.

[0094] Step 304: mutation operation, performing mutation operation according to the mutation probability.

[0095] Step 305: Determine the mileage of the drone formation and calculate the flight distance of each drone formation. If it is greater than the mileage constraint, reallocate the tasks of the drone formation. In the next stage, reduce the tasks or select the nearest tasks.

[0096] Step 306: Calculate the fitness value and input the generated new generation of individuals into the simulated annealing algorithm.

[0097] Step 307: The simulated annealing algorithm continues to perform local search until the maximum number of iterations is reached and the result is output.

[0098] In another exemplary embodiment, Figure 7 As shown, step 205 can be replaced by steps 401 to 406:

[0099] Step 401: First judgment process: Based on the fitness of the initial solution, determine whether the temperature of this iteration has reached the termination temperature or the number of iterations. If so, execute step 402; if not, execute step 403.

[0100] Step 402: Output the optimal drone swarm formation configuration and spectrum resource joint optimization solution.

[0101] Step 403: Generate a neighborhood solution and evaluate the fitness of the neighborhood solution.

[0102] Step 404: Determine whether the fitness of the neighborhood solution is better than the fitness of the initial solution. If so, execute step 405; if not, execute step 406.

[0103] Step 405: accept the neighborhood solution, update the current iteration temperature and the number of iterations, and return to the first judgment process.

[0104] Step 406: With probability e -(fit(R’)-fit(R) / T) Accept the neighborhood solution; where fit(R') is the fitness of the neighborhood solution; fit(R) is the fitness of the initial solution; and T is the current temperature.

[0105] To verify the efficiency and effectiveness of the proposed improved genetic annealing hybrid algorithm, it is compared with heuristic algorithms such as genetic algorithm and simulated annealing algorithm. The simulation also shows the algorithm performance under different task parameters. To make the simulation parameter setting more reasonable, the key parameter settings are shown in Table 1.

[0106] Table 1 Simulation parameter setting table

[0107]

[0108]

[0109] Figure 8 and Fig. 9 The drone quantity planning diagram and frequency selection diagram for drone mission planning with 10 drones assigned to three airport drone formations with mission numbers of 8 and 10 are shown respectively.

[0110] Fig.10 and Fig.11The results of task allocation, UAV number division and bandwidth selection are given when the number of tasks is 8 and 10 respectively. The algorithm not only considers the UAV loss, but also the benefits of executing tasks, so it also reasonably configures the number of UAVs in each UAV formation. Since flight energy consumption is still the main factor affecting the objective function, it can be seen that the UAV formation will still choose to execute the nearest task. However, due to the mileage limit, the tasks that each UAV formation can choose are limited, which also explains the necessity of setting multiple take-off points. The increase in bandwidth will speed up the transmission of information and reduce the hovering loss. Therefore, UAV formation 2 with the largest task volume is allocated the largest bandwidth. The increase in UAVs in the UAV formation will not only increase the completion probability (benefit) of the task, but also increase the energy loss. In the established model, the relationship between the two is reasonably balanced. The final result shows that the number of UAVs allocated to UAV formation 3 has decreased when the task allocation remains unchanged. Because the tasks assigned to UAV formation 1 have changed, the resources it obtains have changed. This also reflects the cooperation of the algorithm in allocating the number of UAVs.

[0111] Fig.12 and Fig.13 The result graphs of different algorithms are given for the number of tasks of 8 and 10, with a total of 10 drones assigned to the drone formation of 3 take-off points. It can be seen that the simulated annealing algorithm is greatly affected by initialization. And the simulated annealing algorithm is easy to fall into the local optimal solution. After the same number of iterations, it quickly converges to a smaller benefit value. Compared with other algorithms, its optimization results are obviously worse. The proposed algorithm, combined with the simulated annealing algorithm, can be seen to converge to a more refined value compared with the traditional genetic algorithm. The simulated annealing algorithm enhances the algorithm's optimization ability and can better jump out of the local optimal solution.

[0112] Fig.14 The benefit result graphs for different numbers of drones under two different weights are given. Regardless of the weight value, as the number of drones increases, the curve first increases to the maximum value and then decreases. The increase in utility reflects the advantage of cooperation, while the decrease in utility reflects that the cost of cooperation is greater than the benefit. In addition, the ξ value indicates the importance of energy consumption. The higher the ξ value, the more energy is consumed when flying in the air. The smaller the ξ value, the smaller the energy consumption of cooperation, so the larger the number of drones corresponding to the peak. The genetic simulated annealing algorithm has a more stable result and can converge to a better value than the simulated annealing algorithm.

[0113] Fig.15The stability comparison of the algorithms under the condition of adding a take-off point with the same resources is given. The existing 30M bandwidth is divided into 3 gears and 4 gears, 10 drones perform 8 tasks, and an additional take-off point (drone formation) is added. The convergence results of the genetic simulated annealing algorithm and the simulated annealing algorithm are compared. It can be clearly seen that because a take-off point is added, the selection of tasks by the drone formation and the drone formation has increased, and its initial value is small, and the result convergence graph is slow. However, because a starting point is added, the path loss is also reduced, and its final benefit result can converge to a larger value. Compared with the traditional simulated annealing algorithm, it converges faster, has better global exploration and local search capabilities, can search for the optimal solution in a large range within the solution space, and with the increase of complexity, the advantages of the algorithm are more obvious. The dispersion of resources does not affect the increase of the final benefit, and the algorithm can reasonably plan tasks and resources according to different starting points, which is effective.

[0114] This application studies the problem of UAV formation mission planning and spectrum resource allocation. By selecting and sequencing tasks, planning the number of UAVs, and selecting bandwidth, the mission benefit function is maximized, and the benefits reasonably balance the relationship between benefits and losses. By utilizing the coupling relationship between the variables to be optimized and the benefits and energy consumption, the problem is transformed into mission network benefit optimization, and a genetic simulated annealing hybrid algorithm is designed. First, the global search capability of the genetic algorithm is used to generate widely distributed candidate solutions, and then the concentrated search characteristics of the simulated annealing algorithm are used to search locally to obtain the final better solution. The proposed algorithm can plan the target perception order, spectrum selection gear, and number of UAV formations for each UAV according to the changes in the mission volume and the setting of the UAV take-off point. The simulation results prove the effectiveness of the proposed algorithm, which can maximize the network benefit function of the mission, balance the relationship between benefits and energy consumption, and the system performance will be better. In future work, the dynamic planning of UAV formations and spectrum resources for mission changes will be considered to cope with complex and changing battlefield environments.

[0115] Based on the same inventive concept, the embodiment of the present application also provides a drone cluster formation configuration and spectrum resource joint optimization system for implementing the above-mentioned drone cluster formation configuration and spectrum resource joint optimization method. The implementation solution for solving the problem provided by the system is similar to the implementation solution recorded in the above-mentioned method, so the specific limitations in one or more drone cluster formation configuration and spectrum resource joint optimization system embodiments provided below can refer to the limitations of the drone cluster formation configuration and spectrum resource joint optimization method above, and will not be repeated here.

[0116] In an exemplary embodiment, a system for configuring a drone swarm formation and jointly optimizing spectrum resources is provided, including:

[0117] The optimization problem model building module is used to dispatch multiple drone formations to perform multiple tasks under the conditions of given spectrum resources and a limited number of drones, and to build an optimization problem model using the coupling relationship between the number of drone formations, target perception order, and spectrum resource allocation; wherein the target perception order is the execution order of each drone formation performing a task; and one target is executed by one drone formation.

[0118] The optimal UAV swarm formation configuration and spectrum resource joint optimization solution output module is used to solve the optimization problem model by using an improved genetic algorithm, and output the optimal UAV swarm formation configuration and spectrum resource joint optimization solution; wherein, the improved genetic algorithm is to embed a simulated annealing algorithm in a genetic algorithm, adjust the crossover operator of the genetic algorithm, and introduce an elite strategy and a dynamically adjustable mutation rate in the mutation operation; adjusting the crossover operator of the genetic algorithm can inherit the characteristics of different parent generations while ensuring the correctness of the offspring, avoiding the omission and duplication of the target; the elite strategy is used to retain the best individuals and directly enter the next generation without going through crossover and mutation operations; the dynamically adjustable mutation rate is used to dynamically adjust the mutation rate according to the operation of the improved genetic algorithm, and the mutation rate is used to explore the solution space and adjust the solution; the simulated annealing algorithm is used to accept poor solutions in the high temperature stage, and gradually converge to the optimal solution from the poor solution in the low temperature stage; the optimal solution is the optimal UAV swarm formation configuration and spectrum resource joint optimization solution.

[0119] In an exemplary embodiment, a computer device is provided, which may be a server or a terminal. The computer device includes a processor, a memory, an input / output interface (I / O for short) and a communication interface. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. 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, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store drone cluster formation configuration and spectrum resource joint optimization data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for configuring a drone cluster formation and jointly optimizing spectrum resources is implemented.

[0120] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the above methods are implemented when the processor executes the computer program.

[0121] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and the computer program implements the above methods when executed by a processor.

[0122] In an exemplary embodiment, a computer program product is provided, including a computer program, and the computer program implements the above methods when executed by a processor.

[0123] 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.

[0124] The present application uses specific examples to illustrate the principles and implementation methods of the present application. The above examples are only used to help understand the method and core ideas of the present application. At the same time, for those skilled in the art, according to the ideas of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present application.

Claims

1. A method for joint optimization of UAV swarm formation configuration and spectrum resources, characterized in that: The method for jointly optimizing the configuration of the UAV swarm formation and spectrum resources includes: Under the conditions of given spectrum resources and limited number of drones, multiple drone formations are dispatched to perform multiple tasks. The optimization problem model is constructed by using the coupling relationship between the number of drone formations, target perception order and spectrum resource allocation; wherein the target perception order is the execution order of each drone formation to perform tasks; one target is executed by one drone formation; An improved genetic algorithm is used to solve the optimization problem model, and an optimal UAV cluster formation configuration and spectrum resource joint optimization solution is output; wherein the improved genetic algorithm is a simulated annealing algorithm embedded in the genetic algorithm, and the crossover operator of the genetic algorithm is adjusted, and an elite strategy and a dynamically adjustable mutation rate are introduced in the mutation operation; the elite strategy is used to retain the best individuals and directly enter the next generation without undergoing crossover and mutation operations; the dynamically adjustable mutation rate is used to dynamically adjust the mutation rate according to the operation of the improved genetic algorithm, and the mutation rate is used to explore the solution space and adjust the solution; the simulated annealing algorithm is used to accept poor solutions in the high temperature stage, and gradually converge to the optimal solution from the poor solutions in the low temperature stage; the optimal solution is the optimal UAV cluster formation configuration and spectrum resource joint optimization solution; the poor solution is a solution with worse quality than the current solution.

2. The method for joint optimization of UAV swarm formation configuration and spectrum resources according to claim 1 is characterized in that: The total distance S of any UAV formation n n for: The total distance traveled by all drone formations is in, Indicates the distance from the airport of UAV formation i to the first selected mission point, Indicates the last mission point to team n's airport π n (0) distance; K n is the total number of tasks assigned to UAV formation n; represents the kth mission point π selected by the UAV formation n n (k) to the k+1th task point π n (k+1) distance; |A n | represents the number of drones in the drone formation n; N is the total number of drones in the drone formation.

3. The method for joint optimization of UAV swarm formation configuration and spectrum resources according to claim 1 is characterized in that: The optimization problem model is: C4: S≤S max C5:|A n |>0 Among them, U is the total utility function; k n is the mission number assigned to formation n; N is the total number of UAV formations; p k is the completion probability of a drone completing task k; |A n | is the number of drones in the drone formation n; V(k) is the value of the task; γ is the weight coefficient of the hovering loss; P(0) is the unmanned power loss during hovering; T(k) is the fixed reconnaissance time of task k, λ k is the amount of information data that needs to be transmitted for task k; W n is the bandwidth of UAV formation n; P m is the transmitting power of the UAV; is the distance from the mission point to the airport of UAV formation n when UAV formation n performs mission k; β0 is the channel power gain at a distance of 1m, σ 2 is the power spectrum density of environmental noise; λ is the weight coefficient of path loss; V mr The flight speed at which the drone cruises the longest distance; is the time for UAV formation n to traverse the mission point; X k,n Perform task k for UAV formation n; n is the sequence number of the UAV formation; K is the total number of tasks; M is the number of UAVs that can be assigned; S max is the maximum mileage of the drone; X n is the order of the formation that performs the task; M is the total number of drones that can be allocated; W is the total bandwidth of the drone formation; C1 means that each task corresponds to a formation; C2 all tasks are performed by drone formations; C3 the sum of the drones in each formation is the number to be allocated; C4 means that the formation mileage must be less than the limit value, C5 means that each formation must be allocated a drone; C6 means that the sum of the bandwidths of each formation is the existing bandwidth; C7 means that the bandwidth of each formation is greater than 0.

4. The method for joint optimization of UAV swarm formation configuration and spectrum resources according to claim 1, characterized in that: The improved genetic algorithm is used to solve the optimization problem model and output the optimal UAV swarm formation configuration and spectrum resource joint optimization solution, which specifically includes: A chromosome is used to initialize the execution task plan of the drone formation to generate the first generation of individuals; a chromosome consists of four lengths. The first chromosome with a length of K represents the perception order of each target, the second chromosome with a length of K represents the tasks assigned to each airport, the third chromosome with a length of N represents the number of drones in each drone formation, and the fourth chromosome with a length of N represents the bandwidth allocated to each drone formation. Introducing the elite strategy, the best individuals are selected from the first generation of individuals and directly used as the next generation of individuals without crossover and mutation operations; The remaining individuals in the first generation are subjected to crossover operation, mutation operation and drone mileage limit in turn to generate the next generation of individuals, and the optimal individuals are combined to generate a new generation of individuals; wherein the drone mileage limit is used to perform feasibility processing on the new schemes generated after the crossover operation and mutation operation, and to determine the mileage of each drone formation to perform the task; Using the new generation of individuals as the initial solution of the simulated annealing algorithm, and evaluating the fitness of the initial solution; Based on the fitness of the initial solution, the new generation of individuals is iteratively trained until a maximum number of iterations is reached, and an optimal UAV cluster formation configuration and spectrum resource joint optimization solution is output.

5. The method for joint optimization of UAV swarm formation configuration and spectrum resources according to claim 4 is characterized in that: Based on the fitness of the initial solution, the new generation of individuals is iteratively trained until the maximum number of iterations is reached, and the optimal UAV swarm formation configuration and spectrum resource joint optimization solution is output, which specifically includes: Based on the fitness of the initial solution, determine whether the temperature of this iteration reaches the termination temperature or whether the maximum number of iterations is reached, and obtain a first determination result; If the first judgment result is yes, output the optimal UAV cluster formation configuration and spectrum resource joint optimization solution; If the first judgment result is no, generating a neighborhood solution and evaluating the fitness of the neighborhood solution; Determine whether the fitness of the neighborhood solution is better than the fitness of the initial solution, and obtain a second determination result; If the second judgment result is no, with probability e -(fit(R’)-fit(R) / T) Accept the neighborhood solution. If the fitness of the new solution is worse than the current solution, then accept the new solution. If the fitness of the new solution is not as good as the current solution, then replace it with probability e. -(fit(R’)-fit(R) / T) Accept the new solution; where fit(R') is the fitness of the neighborhood solution; fit(R) is the fitness of the initial solution; and T is the current temperature.

6. A UAV swarm formation configuration and spectrum resource joint optimization system, characterized in that: The UAV swarm formation configuration and spectrum resource joint optimization system includes: The optimization problem model building module is used to dispatch multiple drone formations to perform multiple tasks under the conditions of given spectrum resources and a limited number of drones, and to build an optimization problem model by using the coupling relationship between the number of drone formations, the target perception order, and the spectrum resource allocation; wherein the target perception order is the execution order of each drone formation to perform tasks; and one target is executed by one drone formation; The optimal UAV swarm formation configuration and spectrum resource joint optimization solution output module is used to solve the optimization problem model by using an improved genetic algorithm, and output the optimal UAV swarm formation configuration and spectrum resource joint optimization solution; wherein, the improved genetic algorithm is to embed a simulated annealing algorithm in the genetic algorithm, adjust the crossover operator of the genetic algorithm, and introduce an elite strategy and a dynamically adjustable mutation rate in the mutation operation; adjusting the crossover operator of the genetic algorithm can inherit the characteristics of different parent generations while ensuring the correctness of the offspring; the elite strategy is used to retain the best individuals and directly enter the next generation without crossover and mutation operations; the dynamically adjustable mutation rate is used to dynamically adjust the mutation rate according to the operation of the improved genetic algorithm, and the mutation rate is used to explore the solution space and adjust the solution; the simulated annealing algorithm is used to accept poor solutions in the high temperature stage, and gradually converge from the poor solutions to the optimal solution in the low temperature stage; the optimal solution is the optimal UAV swarm formation configuration and spectrum resource joint optimization solution.

7. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for joint optimization of drone cluster formation configuration and spectrum resources as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for jointly optimizing the configuration of the drone cluster formation and spectrum resources described in any one of claims 1 to 5 is implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for jointly optimizing the configuration of the drone cluster formation and spectrum resources described in any one of claims 1 to 5 is implemented.

Images