论文部分内容阅读
PID控制由于其结构简单、稳定性好而被广泛应用,然而在实际的工业过程中,许多被控过程机理复杂,具有高度非线性。在实际PID控制中就要求,PID参数不仅整定不依赖对象数学模型,并且能在线调整,以满足实时控制的要求。对角回归神经网络是一种具有反馈回路的动态神经网络,本文提出了一种基于对角回归神经网络的PID控制器结构,利用对角回归型神经网络辨识控制量权值调整中的未知Jacobian信息,在线整定PID控制器参数。仿真实验中,将此法与基于径向基函数神经网络PID控制效果进行对比,结果显示基于对角回归型神经网络的自适应PID控制器在抗干扰、设定值跟踪动态响应和鲁棒性上都有较明显的改善。
PID control is widely used due to its simple structure and good stability. However, in actual industrial processes, many controlled process mechanisms are complex and highly non-linear. In the actual PID control requirements, PID parameters not only set does not rely on the object mathematical model, and can be adjusted online to meet the requirements of real-time control. The diagonal regression neural network is a kind of dynamic neural network with feedback loop. In this paper, a PID controller structure based on diagonal regression neural network is proposed. By using the diagonal regression neural network, the unknown Jacobian Information, online tuning PID controller parameters. In the simulation experiment, this method is compared with PID control based on radial basis function neural network. The results show that adaptive PID controller based on diagonal regression neural network has good performance in anti-jamming, setting tracking dynamic response and robustness On the more obvious improvement.