摘要
针对飞控系统的可重构设计问题,提出了解耦控制与自适应逆控制(AdaptiveInverseControl)相结合的控制策略。首先采用线性动态逆方法设计了某型飞机的滚转角控制回路,实现飞机滚转与侧滑之间的解耦;其次,给出了具有可变增益的基于ε滤波的LMS算法,在系统解耦的基础上构建了自适应逆控制系统,从而实现系统在舵机故障下的重构控制。仿真结果表明,这种基于改进的ε滤波可变增益LMS算法的自适应逆可重构飞行控制系统具有良好的系统故障恢复功能,表明了其在飞控系统重构中的有效性。
A control strategy of decoupling control combined with adaptive inverse control is presented for the reconfigurable control problem of multi-input multi-output systems. First, the roll angle control loop of a fighter aircraft is designed with linear dynamic inversion method, so that the sideslip angle and roll rate of the aircraft's lateral dynamics can be decoupled. Then, based on the designed decoupled system, an adaptive inverse control system is constructed using the filtered-ε LMS(Least Mean Square) algorithm to reconfigure the damaged system in case of failures of actuators. To avoid the adaptation process of the plant inverse model in this system, a modified filtered-ε structure is introduced and thus only an approximate plant inverse model is needed; this simplifies the design very much. In general, a large gain makes the LMS algorithm have a better tracking ability but also makes it sensitive to noise leading to a non-smooth steady-state performance. On the other hand, a small gain makes the algorithms less sensitive to noise but results in a poor tracking ability. A variable gain LMS algorithm is then introduced to satisfy both a fast tracking ability and a smooth steady-state requirement. Simulation results show that the reconfigured roll angle response of the aircraft tracks the step series command input well after only about 5 seconds. The simulation results also show that the adaptive inverse reconfigurable flight control system based on the modified filtered-ε variable gain LMS algorithm works well in case of actuator failures.
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2005年第3期332-336,共5页
Journal of Northwestern Polytechnical University
关键词
可重构控制
解耦控制
可变增益LMS算法
自适应逆控制
reconfigurable control, decoupling control, variable gain LMS(Least Mean Square) algorithm, adaptive inverse control