摘要
BP神经网络已被广泛应用于PID控制器的优化调参,但这种调参方法具有收敛速度慢、学习时间长、连接权重初值为随机值、易于陷入局部极小等缺点。提出了一种不同于用BP网络调整PID参数的新的融合方法:PID神经网络控制器(PIDNN);该控制器不仅能克服以上缺点,而且具有很好的鲁棒性。对PIDNN在某无人机姿态控制系统的应用进行了仿真研究,仿真结果表明该控制器能够大大地改善姿态控制系统性能。
The neural network with BP algorithm has these disadvantages, such as slow convergence rate, long learn time, local optimization and random value for the initial weight of the net, although it has been widely applied to regulate the parameters of PID controller. A new type of controller named PIDNN, which coalesces traditional PID and neural network together, is introduced in this paper. It overcomes these shortcomings mentioned and has better robustness. Then it is used in the flight control system of UAV, the simulation results show that it improves the performance of the system effectively.
出处
《航空计算技术》
2008年第2期6-9,共4页
Aeronautical Computing Technique
基金
教育部新世纪人才支持基金(NCET-05-0357)
