一种基于在线辨识和专家系统的飞行器智能姿态控制方法

针对未来战场形势对攻击型飞行器提出的大包线飞行、大不确定、主动容错控制问题,提出了一种基于在线辨识和专家系统的智能姿态控制方法。该方法通过在线估计飞行器的时域、频域关键参数,驱动专家系统实时计算并调整控制器参数;结合时域、频域辨识结果诊断、定位故障,实现故障隔离、控制策略调整。该方法综合了飞行器结构、气动试验数据和专家知识,通过在线辨识和估计丰富了专家系统推理计算所需信息,促进了专家系统的在线应用,是一种近期可实现的智能控制方法。仿真结果表明,采用该方法,姿态发散概率由无容错控制时的45%降低到容错控制时的0%,姿态控制任务完成率100%。

An Intelligent Attitude Control Method for Flight Vehicle Based on On-line Identification and Expert System

To meet the requirement of future warfare for attack flight vehicle in terms of large envelope flight control, high uncertainty and active fault-tolerant control(FTC), an intelligent attitude control method based on on-line identification and expert system is proposed. In this method, the realtime calculation and adjustment of the controller parameters can be achieved by the expert system with the online estimation of the crucial frequency-domain or time-domain parameters by using the frequency-domain or time-domain identification results, the fault diagnosis and location can be accomplished, and the fault isolation and control strategy can be adjusted. This method, which integrates the flight vehicle structure, aero-dynamical test data and expert knowledge, enriches the information for reasoning calculation of expert system through online identification and estimation, and promotes the application of expert system on line, is an intelligent control method that can be realized in the near future. The simulation results show that, in fault tolerant control simulation, the probability of attitude divergence is reduced from 45%(without FTC) to 0% (under FTC), and the completion rate of attitude control task is 100%.