一种快速非线性模型预测控制律研究

Research on a Fast Nonlinear Model Predictive Control Law

为滑翔增程弹设计了一种滑翔控制律,在考虑其非线性、状态约束和控制舵偏角幅值约束的情况下,针对滑翔控制过程中采样周期短,系统动态变化快的特点,采用直接多重打靶法和实时迭代策略来缩短非线性模型预测控制中递推优化的计算时间,并应用预先优化策略避免反馈控制延迟。仿真结果表明,该控制器的控制律求解时间小于采样周期,能满足滑翔增程弹滑翔控制要求,控制效果良好。研究结果可为同类弹箭非线性模型预测控制律设计提供参考。

Taking nonlinear characteristics, state constraints and magnitude limits of control surface deflections into account, a gliding control law was designed for glide range-extended projectile. For the characteristics of short sampling time and fast system dynamics during the glide control process, the direct multiple shooting method and the real-time iteration scheme were employed to reduce the runtime needed for the receding optimization of nonlinear model predictive control, and the advanced-step scheme was also used to avoid the feedback control delay. The simulation results show that the runtime for solving control law is less than sampling period and the controller could satisfy the control requirement of glide rangeextended projectile. The research results provide an important reference for nonlinear model predictive control law design of similar projectiles and rockets.