基于神经网络自适应PID的无人机编队避障飞行控制研究

Research on UAV Formation Obstacle Avoidance Flight Based on Neural Network Adaptive PID Control

针对无人机编队避障飞行控制难题,研究了无人机编队避障航迹规划与智能控制技术。首先,提出一种基于改进人工势场法的无人机编队航迹规划算法,利用改进势场函数和引入“随机波动”法等手段,解决了传统人工势场法用于无人机编队航迹规划时遇到的无法到达目标点以及局部最小值问题,并提升了传统算法航迹规划的快速性和鲁棒性。其次,设计了一种新的BP神经网络辅助的自适应PID无人机编队智能控制算法,通过利用神经网络的在线学习能力,实现了PID参数的优化整定,提高了现有PID算法的计算精度。最后,通过无人机编队避障飞行控制仿真实验,验证了提出方法的有效性,基于NN-PID控制律的编队控制器能够更好地对无人机编队进行有效的控制。

Aiming at the difficult problem of UAV formation obstacle avoidance flight control, the UAV formation trajectory planning and intelligent control technology has been carried out. First, an improved artificial potential field UAV formation trajectory planning algorithm is proposed, which solves the problem of unable to reach the target point and local minimum encountered of traditional artificial potential field method for UAV formation trajectory planning by improving the potential field function and introducing the “random fluctuation” method. And improve the speed and robustness of traditional algorithm trajectory planning. Secondly, a new neural network-assisted adaptive PID UAV formation intelligent control algorithm is designed. By using the online learning ability of the neural network, the optimization of PID parameters is realized, and the calculation accuracy of the existing PID algorithm is improved. Finally, through the UAV formation obstacle avoidance flight control simulation experiment, the effectiveness of the method proposed in this paper is verified.