节点文献
基于极限学习机的有限时间自适应姿态控制
Extreme Learning Machine-Based Finite-Time Adaptive Attitude Control
【摘要】 对存在转动惯量未知摄动和角速度有界约束的航天器姿态控制问题,设计了一种基于非奇异终端滑模面的有限时间姿态跟踪控制器。预设有界虚拟角速度作为外环系统的控制输入,使姿态角快速收敛到期望姿态。将外环虚拟角速度作为内环系统的期望轨迹,采用极限学习机(Extreme Learning Machine,ELM)在线估计包含未知转动惯量的部分,并基于非奇异终端滑模面设计了一种自适应姿态控制器,使内环输出角速度在有限时间内精确跟踪预设虚拟角速度,满足了角速度有界约束。理论分析了闭环系统的全局渐近稳定性。数值仿真了在轨抓捕非合作目标过程中的航天器姿态控制问题,验证了本文控制器的合理性和姿态控制器的优良品质。
【Abstract】 A novel non-singular terminal sliding mode based dual-loop attitude tracking controller is provided for a rigid spacecraft due to the uncertain rotation inertia,external disturbances and bound angular velocity. The attitude dynamics equation is decomposed into the inner and outer loop system. A bound virtual angular velocity is preplanned as the control input of outer loop system to ensure the attitude converges rapidly. The preset virtual angular velocity is regard as the desired trajectory of inner loop system. A adaptive sliding mode attitude controller is designed for inner loop system,in which an ELM is constructed to estimate the part of uncertain rotational inertia,and a non-singular terminal sliding mode surface is provided to ensure the tracking error between real and virtual angular velocities,which converges to zero in the finitetime. The constraint of bounded angular velocity is fulfilled,and asymptotical stability of the close-loop system is theoretically analyzed. The attitude control problem of spacecraft capturing non-cooperative target is simulated by numerical simulation and the rationality and robustness of controller are verified.
【Key words】 Attitude control; Finite-time; Terminal sliding mode; Dual-loop tracking control; Extreme learning machine;
- 【文献出处】 航天控制 ,Aerospace Control , 编辑部邮箱 ,2018年05期
- 【分类号】V448.2
- 【被引频次】2
- 【下载频次】157