一种智能反步控制方法及其应用

An Optimal Intelligent Backstepping Control Method for Reaction-Jet Missile

针对反作用射流导弹的非线性和参数不确定性严重的特点,提出一种鲁棒自适应反步智能控制(Backstepping)方法。通过非特定的Lyapunov函数,采用遗传算法优化Backstepping镇定系数,得到适当的函数降低了当镇定系数不当时跟踪误差对控制律的影响。结合模糊小脑模型(FC-MAC)神经网络,推出系统参数不确定时的鲁棒控制律,克服了传统自适应反步法对模型不确定性要求线性有界问题。仿真结果表明,这种结合Lyapunov函数选择和参数优化的Backstepping方法鲁棒性强,受参数变化影响小。

In the full paper, we explain our control method in detail; in this abstract we just add some pertinent remarks to listing the two topics of explanation： （1）combining RCS （reaction jet control system） with aerodynamic surfaces and （2） the design and analysis of nonlinear robust backstepping attitude controller; in topic 1, eqs. （2a）, （2b） and （2c） in the full paper include the terms for both RCS and aerodynamic control surfaces; the three subtopics of topic 2 are the design of backstepping control law （subtopic 2. 1）, the adjustment of parameters （subtopic 2. 2） and the adaptive control law for uncertain system（subtopic 2.3）; in subtopic 2. 1,we choose a form of universal Lyapunov function and then, by analyzing the system dynamics properties, we use GA （genetic algorithm） to optimize the regression coefficients so as to suppress the effect of unavoidable unmatched backstepping coefficients; in subtopic 2.3, we use FCMAC （fuzzy cerebellar model articulation controller） to improve system robustness so as to deal with system uncertainties and to make unnecessary the required knowledge of model uncertainty bound of traditional adaptive backstepping. Simulations results, given in Figs. 2 through 6 in the full paper, show preliminarily that our optimal intelligent backstepping control method is feasible for missile autopilot design.