具有自适应噪声边界的Tube可达集鲁棒预测控制

Tube-reachable set-based robust model predictive control with adaptive disturbances boundaries

针对过程噪声设定边界与真实噪声边界失配的有界干扰离散线性不确定系统,提出一种具有自适应噪声边界的Tube可达集鲁棒模型预测控制方法.首先,该算法引入基于MIT规则的自适应集员滤波在线估计系统状态和噪声边界.其次,基于估计值,通过迭代自适应集员滤波的时间更新部分计算出预测时域内闭环不确定系统状态的可达集.最后,用可达集代替不变集并根据Tube鲁棒模型预测控制策略,给出了实际不确定系统的控制律,确保系统状态鲁棒渐近稳定,并收敛于终端干扰不变集.仿真结果验证了该控制方法的有效性.

For a discrete-time linear uncertain system with bounded disturbances but mismatch between the real process noise boundary and its setting boundary, we present a tube-reachable set-based robust model predictive control （RMPC） with adaptive disturbances boundaries. Firstly, an MIT-based adaptive set-membership filter （MIT-ASMF） is introduced to estimate the states and the boundaries of the process noise in real time. Secondly, based on the estimated results, the reachable sets of the local uncertain closed-loop system are calculated by propagating the prediction step of the ASMF along the prediction horizon. Finally, by replacing the invariant set with previously determined reachable sets in the tube-based RMPC formulation, we determine the control law for the real uncertain system to guarantee the real uncertain system of robustly asymptotic stability and the convergence to the terminal disturbance invariant set. The simulation results validate the efficacy of the proposed method.