基于参数估计误差的鲁棒自适应律设计及验证

针对传统自适应控制系统设计的自适应律参数收敛慢进而影响控制系统瞬态性能的问题,研究一类新的基于参数估计误差修正的鲁棒自适应律设计.首先引入滤波操作给出参数估计误差的提取方法,构建出含参数估计误差修正项的自适应律,进而将该自适应律用于控制器设计和分析中,可同时实现控制误差和参数估计误差指数收敛.对比分析了几类传统自适应律和所提出自适应律的收敛性和鲁棒性,并给出了保证参数收敛所需持续激励条件的一种直观、简便的在线判别方法.数值仿真及基于自制三自由度直升机系统俯仰轴实验结果表明,基于参数误差修正的自适应律及控制器可得到优于传统自适应方法的跟踪控制和参数估计性能.

Parameter estimation error based robust adaptive law design and experiments

To address the sluggish parameter estimation and control performance of conventional adaptive control methods, this paper presents a new robust adaptive law based on the parameter estimation error. The information of parameter estimation error is first derived via filter operations, which is used as a new leakage term in the adaptive law design. The proposed adaptive law is further incorporated into the adaptive control system, which can retain exponential convergence of the tracking control error and the parameter estimation simultaneously. Moreover, comparisons to other classical adaptive laws are provided with respect to the parameter estimation convergence and robustness property. A simple and intuitive approach to online test the required persistent excitation (PE) condition is suggested. Simulation and experimental results based on a small 3 DOF-helicopter show that the newly suggested adaptive method can achieve better control and estimation performance than other conventional adaptive schemes.