基于宽度学习的航天器编队分布式姿态协同容错控制  

Distributed Attitude Cooperative Fault-Tolerant Control of Spacecraft Formations Based on Broad Learning System

作　　者：俞鑫丽 易辉[1] YU Xinli;YI Hui(Nanjing Tech University,Nanjing 211816,China)

机构地区：[1]南京工业大学,南京211816

出　　处：《航天控制》2024年第3期54-60,共7页Aerospace Control

基　　金：国家自然科学优秀青年科学基金项目(61922042);国家自然科学基金重点国际(地区)合作研究项目(62020106003);江苏省前沿引领技术基础研究重大项目(2022050029)。

摘　　要：针对具有惯性不确定和执行器故障的航天器编队姿态协同控制问题,利用宽度学习系统的逼近特性对系统的广义扰动和执行器故障进行估计,同时采用迟滞量化器对控制力矩信号进行量化,以降低对通信速率的要求,并减少抖振现象。在此基础上,提出一种基于模型预测控制和快速非奇异积分终端滑模的复合结构容错控制器。利用代数图论和Lyapunov理论分析了闭环姿态系统的稳定性。最后,通过仿真验证了本文提出的控制方法与现有方法相比的优越性。According to inertial uncertainties and multiple types of actuator faults of spacecraft formation attitude cooperative control,the generalized disturbances and actuator faults of the formation system are estimated by using the approximation property of the broad learning system,in the meantime,the hysteresis quantizer is also used to quantize the control torque signal,thereby reducing the communication speed requirements and jitter phenomenon.On this basis,a composite structure fault-tolerant controller based on model predictive control and fast non-singular integral terminal sliding mode is proposed.The stability of the closed-loop attitude system is analyzed by involving algebraic graph theory and Lyapunov theory.Finally,simulation results show the superiority of the proposed control method compared with existing methods.

关 键 词：航天器编队 姿态跟踪 容错控制 模型预测控制 输入量化 宽度学习系统 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]