一种鲁棒输出反馈线性化方法在导弹自动驾驶仪设计中的应用研究

针对具有多模型的HAVE DASHⅡ导弹对象，利用一种结构简单的鲁棒输出反馈线性化方法，为其设计了全弹道自动驾驶仪控制系统。设计中仅需要少量的模型集信息，并不依赖于某一具体子系统，对整个模型集具有通用性。由于不再需要设计多个控制器，因此避免了常规多模型切换摔制中抖动的产生。该方法利用模型估计器实时估计系统动态特性，只要参数选得合适，能无偏差地估计出模型集中每个系统动态特性，有很强的鲁棒性。此外。还利用了高增益观测器来获取控制律中所需的导数信息，使得整个闭环系统仪需要输出量反馈，降低该方法的工程实现难度。仿真结果表明，所设计的自动驾驶仪系统能够控制模型集中所有飞行条件下的状态准确快速地到达期望值上。

Application on a Robust Output Feedback Linearization Approach to Missile Autopilot Design

This paper investigates autopilot design for HAVE DASH II missile using a kind of robust output feedback linearization approach with simple structure. Only a little multi-model information was needed. The design of this controller did not depend on any of subsystem. Therefore, it had generality for muhi-model. Because of single robust controller, the switching chattering in traditional multi-model control was avoided. Using model estimator with appropriate parameters, the control law could give the entire dynamics of multi-model without any estimation error. In addition, high gain observers were applied to obtain derivatives in control law. Then in the closed-loop system, only the output variables were used for feedback, This reduced the engineering difficulty. Simulation results are presented to show that, the autopilot designed can ensure states at all flight conditions to their target precisely and quickly.