Research On Intelligent Scheduling Algorithm Of Unmanned Aircraft Cluster Based On Hardware Acceleration

With the improvement of UAV operation autonomy and multi-task requirements,UAV has developed from stand-alone operation to cluster operation.However,due to the complex working environment,unbalanced load,and insufficient group coordination of swarm UAVs,the efficiency of cluster task execution is low,and it is difficult to cope with the application field of dynamic changes in the task environment.high demands.As an effective way to solve this problem,UAV swarm task scheduling technology has attracted the attention of researchers.The essence of UAV swarm task scheduling is a complex multi-constraint and multicombination optimization problem.The swarm-based meta-heuristic swarm intelligence algorithm provides a solution for UAV swarm task scheduling because of its good global optimization ability and robustness.Based on the analysis of the domestic and foreign status quo of UAV swarm task scheduling,this paper focuses on the optimization of swarm UAV task scheduling strategy and the hardware acceleration method of scheduling algorithm.The main research contents include:(1)The mathematical model of UAV swarm task scheduling and the construction of scheduling decision-making benefit evaluation function.This paper conducts a comprehensive analysis of the drone’s flight speed,maximum flight distance,weapons carried,enemy target attack range,distance from the drone,and enemy target value to balance the performance load of UAVs and enemy targets,decompose the task classification and constraints of UAV groups,constructing a mathematical model of UAV swarm task scheduling and a scheduling decision-making benefit evaluation function.After testing,the scheduling model and scheduling decision-making benefit evaluation function can correctly reflect the advantages and disadvantages of the benefits of UAVs performing scheduling tasks.(2)Design of UAV swarm task scheduling method based on wolf pack algorithm.This paper introduces benchmark functions to analyze and compare the performance of commonly used UAV swarm intelligent scheduling optimization algorithms such as particle swarm algorithm,genetic algorithm,artificial fish swarm algorithm,and wolf swarm algorithm,and determines to use the wolf swarm algorithm to solve the UAV swarm task scheduling decision optimization problem constructed in this paper.The wolf pack algorithm is used to solve the task scheduling decision-making optimization problem to improve the efficiency of UAV swarm task execution.The test results show that this method can correctly realize the function of optimal task scheduling decision-making optimization and has high stability.(3)Design of hardware acceleration method for task scheduling algorithm.In order to improve the real-time performance of the scheduling algorithm,this paper proposes a hardware acceleration method for the task scheduling algorithm based on the wolf pack algorithm.Using FPGA to implement and test it in hardware,the results show that,using the same test data,the average calculation time of the scheduling algorithm after hardware acceleration design is about 850 ms under the 100 Mhz system clock,and the average software calculation time is about 15 s.The test results prove that the calculation speed of the hardware acceleration implementation method is about 18 times higher than that of the software implementation.