Research on Emergency Drone Flight Control System Based on Improved Near End Strategy Optimization Algorithm Control
In order to further improve the flight control effectiveness of emergency drones during mission execution,a dimension pruning technique is proposed to optimize and solve the zero gradient problem of the Near End Policy Optimization(PPO)algorithm.While maintaining good sampling efficiency,the convergence rate is accelerated,thereby improving control performance.The simulation test results show that the accuracy of the improved PPO algorithm is greater than 90％at different iterations,with a maximum accuracy of 92％.However,the accuracy of the k-NN algorithm fluctuates to some extent at different iterations,with a maxi-mum accuracy of 90％and a minimum accuracy of only 80％.And the total computational time cost of im-proving the PPO algorithm and PPO algorithm is basically the same,both of which are 1 932.4 seconds.But improving the PPO algorithm can make the loss value converge faster during the training process.
万方数据
维普
