具有磁滞输入的可调金属切削系统鲁棒自适应动态面控制

Robust adaptive dynamic surface control for adjustable metal cutting system with hysteresis input

针对既有时滞环节又存在磁滞输入的可调金属切削系统,提出了一种改进的自适应动态面控制方法,其特点为:1)设计了带有跟踪误差性能指标函数的鲁棒自适应动态面控制算法,并结合神经网络,使其能够保证系统的跟踪误差及其过渡过程在预先任意给定的范围内;2)克服了反推控制方案中的"微分爆炸"问题,简化控制器结构;3)估计神经网络权值向量的范数而不是估计权值向量,极大地减少系统的计算负担,便于实时控制.仿真结果验证了该控制方法的有效性.

We propose an improved control method for the adjustable metal cutting system with hysteresis input and time-delay.The main features are：1） it fuses the performance function into robust adaptive dynamic surface control,combines the neural network,ensures the system tracking error and the transient process to stay in a pre-specified region; 2） it eliminates the explosion of differentiation inherent in the backstepping control,simplifying the structure of the system controller; 3） it estimates the weight vector norm of the neural network instead of estimating the weight vector,thus greatly reduces the computational burden and makes the controller more convenient for real-time applications.Simulation results indicate that the proposed scheme is effective for all unknown time-varying system with hysteresis input,such as the adjustable metal cutting system we mentioned.