阶梯式有约束非线性预测控制及其应用

Stair-like Constrained Nonlinear Model Predictive Control and Its Application

非线性预测控制需要在每个采样周期内求解非凸非线性规划。针对在线求解的计算复杂性,提出基于阶梯控制的非线性预测控制算法,在每个采样周期,采用阶梯式控制策略,精确求解当前控制量,减少非线性规划决策变量的维数,降低了在线计算量,减少计算时间,提高了算法的快速性。仿真和实控结果表明,本算法具有良好的可行性和控制效果。

Nonlinear model predictive control needs to solve non-convex nonlinear programming problem per sampling time. According to the complexity of online computation, a new algorithm of nonlinear model predictive control based on stair-like control was proposed. It approximated the future control sequences by a set of stair-like controllers and only computed the current control action per sampling time. The method decreased the number of optimisable variable, so it could greatly reduce the online computation burden and improve the implementation of the algorithm. Simulation and experimental results prove the effectiveness of the proposed controller.