自适应比例–积分H2滑模观测器设计

Adaptive proportional-integral H2 sliding mode observer design

中文关键词  滑模观测器  径向基函数网络  自适应控制  区域极点配置  鲁棒性

英文关键词  sliding mode observer  radial basis function networks  adaptive control  regional pole assignment  robustness

中文摘要

      传统的龙伯格观测器的观测精度极易受到未知外部扰动的影响. 为了解决这个问题, 本文设计了一种基于径向基神经网络的自适应比例–积分H2滑模观测器, 实现了参数不确定性和外部扰动下非线性系统的鲁棒确切估计. 首先, 利用径向基神经网络自适应逼近系统模型的复杂非线性项; 其次, 设计基于误差的线性滑模面, 将比例–积分滑模项注入观测器中, 使得滑模动态在有限时间内收敛于滑模面, 实现对外部扰动和系统模型非线性的完全补偿; 最后, 基于H2次优控制和区域极点配置, 提出观测器参数自整定方法. 通过对单连杆机器人的仿真结果表明, 该方法能够保证非线性系统具有较好的鲁棒性和自适应性.

英文摘要

      The observation accuracy of traditional Luenberger observer is easily affected by unknown external disturbance. To solve this problem, an adaptive proportional-integral H2 sliding mode observer is designed in this paper, which achieves robust exact estimation of nonlinear systems with parameter uncertainties and external disturbances. Firstly, the radial basis function neural network is used to approach the complex nonlinear terms of the system model. Secondly, a linear sliding mode surface based on error is designed, and the proportional integral sliding mode term is injected into the observer, so that the sliding mode dynamic converges to the sliding mode surface in finite time, and the nonlinear compensation of external disturbance and system model is realized completely. Finally, an observer parameter self-tuning method is proposed based on the H2 suboptimal control and regional pole assignment. The simulation results of a single-link robot verify the proposed method can ensure the robustness and adaptability of the nonlinear system.