电液伺服结构试验系统的神经网络快速鲁棒跟踪控制*

Neural Network Based Fast and Robust Tracking Control for Electro-Hydraulic Servo Structural Testing System

中文关键词  结构试验系统  神经网络控制  在线学习  鲁棒性

英文关键词  structural testing system  neural network control  on-line leaming  robustness

中文摘要

      针对结构试验系统的非线性和不确定性特性，提出一种神经网络并行自学习跟踪控制器，在满足试验系统实时性要求的条件下，通过神经网络在线建模和虚拟学习做到了控制器的在线自适应设计，并解决了实时训练样本不足的问题，实例仿真表明设计的控制系统具有良好的波形再现能力.

英文摘要

      A parallel self-learning tracking controller based on neural networks,with respect to the complex nonlinearties and uncer-tainties of structural testing system,is presented in this paper.By introducing a neural network to model the controlled system and using virtu-al leaming,the controller can be adaptively designed on-line with no need of the real-time measured training data.Meantime, the real-time processing ability of the neural network. control for complex systems is assured. Simulation results of a real control system show it has good tracking ability.