基于特征模型的高超声速飞行器再入多约束预测控制

Predictive Control Based on Characteristic Model for Reentry Hypersonic Vehicle With Multiple Constraints

针对存在输入和输出多约束的高超声速飞行器再入控制问题,提出了一种基于特征模型的鲁棒预测控制方法。对耦合的非线性再入动力学采用低阶线性特征模型简化,设计鲁棒预测控制器保证满足再入控制的多约束。基于特征模型的低阶等价性,将再入动力学的输入输出动态特性用线性时变的特征模型描述,并将三通道姿态耦合和干扰作为广义复合干扰。将灰色理论中累加求和的思想引入到特征模型参数辨识算法中,弱化了再入随机噪声的影响,通过灰色辨识方法在线估计特征模型的时变参数和干扰。设计一种基于线性矩阵不等式(LMI)滚动优化技术的H_2/H_∞鲁棒预测控制律与基于灰色辨识的补偿器共同构成复合控制器,保证了整个闭环系统的稳定,满足输入和输出约束。数据仿真验证了算法的有效性。

In order to study the reentry control problem of hypersonic vehicle with multiple constraints, a novel robust predictive-control-method was proposed based on the characteristic model. The low-order linear characteristic model was adopted to simplify the coupled nonlinear reentry dynamics, and the robust-predictive-controller was designed to satisfy the multiple constraints in the reentry control. Based on the low-order equivalence of characteristic model, the input and output dynamic-characteristic of the reentry dynamics was modeled by linear time-varying model, and the three-channel coupling and external disturbance were considered as a generalized composite disturbance. The accumulative sum idea in the grey theory was introduced into the parameter identification of characteristic model, and the effect of random noise factor was reduced, and the regularity of identification data was strengthened. The time- varying parameters and the disturbance were on-line estimated based on grey identification. The mixed H2/H~ robust predictive control-law was proposed by linear matrix inequality（LMI） receding horizon opti- mization technique, and the compensator of grey identification was feedforward compensated. The closed- loop system is stable, which satisfies the input and state constraints. The validity of the algorithm was verified by simulation results.