论文部分内容阅读
针对复杂非线性系统的控制问题,采用数据驱动的控制策略,将具有本质自适应能力的即时学习算法与具有强鲁棒性的滑模预测控制相结合,设计了一种基于即时学习的滑模预测(LL-SMPC)控制方法.该方法在在线局部建模的基础上,采用滑模预测控制律求取最优控制量,具有较强的自适应和抗干扰能力,并采用分层递阶搜索策略,避免了求解Diophantine方程,有效减少了计算量,提高了搜索效率.仿真结果验证了所提出算法的有效性.
Aiming at the control problem of complex nonlinear systems, a data-driven control strategy is adopted to combine the real-time adaptive learning algorithm with sliding mode predictive control with strong robustness. A sliding mode based on instant learning (LL-SMPC) control method based on on-line local modeling, the method is based on sliding mode predictive control law to get the optimal control quantity, which has strong self-adapting and anti-jamming ability. Search strategy, which avoids solving the Diophantine equation, reduces the computational complexity and improves the search efficiency. The simulation results verify the effectiveness of the proposed algorithm.