模型参考鲁棒控制的重新设计与仿真

提出了一种新的模型参考鲁棒控制方案。分析了以往方案在被控对象相对阶大于1时跟踪性能不理想的原因,通过对误差方程的变换引入新的Lyapunov函数,由此重新设计了其Backstepping控制律。可以证明:即使系统参数未知且存在未建模动态和非线性因素干扰,新方案在保证闭环系统稳定性的同时,可显著改善系统的跟踪性能并降低控制信号的幅度。分别采用新方法、以往的模型参考鲁棒控制和传统的零极点配置方法,对某自动驾驶仪俯仰保持模态控制律设计进行数值仿真对比,结果表明:无论从跟踪性能和控制信号性能来看,都以新方案为最优。

Redesign and Simulation of Model Reference Robust Control

A new model reference robust control(MRRC) scheme was proposed.The earlier design method,the tracking performance was shown,in general,unacceptable.Then,by transforming the error equation,a new Lyapunov function was introduced and a new control law with backstepping method was designed.It can be proved that by using the scheme,even with the existence of plant uncertainty,unmodeled dynamics and non-linear disturbance,both the tracking performance and the control effort can be improved significantly.Finally,the new MRRC algorithm was applied to a pitch-attitude-hold autopilot design.In contrast to the earlier MRRC method and the classical zero-pole placement method,the new algorithm achieves the best performance.