Trajectory Planning And Task Allocation Of UAV In Battlefield

UAV has the obvious advantages of long flight time and good maneuverability,and they are especially suitable for actual combat needs such as precise reconnaissance and strike,large overload,and high-speed flight.UAV trajectory planning and task allocation are necessary conditions for completing combat missions in a battlefield environment.This thesis fully considers the characteristics of information support,complex terrain,multiple threat types,large number of missions,and high timeliness requirements in the actual battlefield and optimizes the shortcomings of the existing trajectory planning and task allocation algorithms,such as insufficient intelligence utilization,incomplete threat considerations,impractical combat style,and insufficient convergence speed to improve UAV autonomy.The main research contents are as follows:The existing UAV route planning algorithm is mainly based on two-dimensional terrain,which cannot meet the three-dimensional terrain and multi-threat environment,and the prior information is not fully utilized and the convergence speed is slow.To solve this problem,firstly,a real battlefield environment model containing threat source features is constructed,and the three-dimensional environment model is reduced in dimension through the reward and punishment matrix in the Q-learning algorithm to avoid "dimensional disaster".Secondly,the use of intelligence information formulates the initial Q table information in the Q learning algorithm,the priority of the action selection mechanism and the feedback value,which provides a basis for the flight direction of the drone.Finally,for the situation where the battlefield environment is prone to mutation,an emergency route planning algorithm based on the retreat action selection mechanism is designed,which increases the survival rate of the drone.The simulation results show that the UAV battlefield route planning based on the improved Q learning algorithm can effectively use the a priori information to reduce the environmental dimension,improve the convergence rate and the adaptability of the drone to sudden environmental changes,and enhance Autonomy.In order to fully improve the autonomy of the UAV and closely integrate the combat style,the task assignment of the UAV is also the key.At first,the "drones bee colony" that meets the requirements of the mission is screened by constructing a reverse auction algorithm to screen the range of available drones is clarified for the next optimal attack.Next the fitness function of the particle swarm algorithm is constructed,and an improved particle swarm algorithm with variable weight factor and cognitive coefficient is proposed to realize the "drones bee swarm" sequential attack mode.The simulation results show that the proposed UAV task allocation based on auction algorithm and improved particle swarm algorithm adapts to the new combat style,improves the convergence rate and mission revenue,and increases the autonomy of the UAV.