Model predictive steering control law for double gimbal scissored-pair control moment gyros

Model predictive steering control law for double gimbal scissored-pair control moment gyros
复制标题

DOI:
10.1016/j.actaastro.2021.03.023
复制
发表时间:
2021-04
期刊:
影响因子:
3.5
通讯作者:
H. Kojima;Reiji Nakamura;S. Keshtkar
H. Kojima;Reiji Nakamura;S. Keshtkar
中科院分区:
工程技术3区
文献类型:
--
作者:
H. Kojima;Reiji Nakamura;S. Keshtkar

文献摘要

相似文献

为了克服控制力矩陀螺仪(CMG)的奇点问题,同时提供轻量级、敏捷且简单的动力系统,最近提出了一种称为双万向剪式对控制力矩陀螺仪(DGSPCMG)的CMG系统。 DGSPCMG 是一种混合机构,结合了剪式 CMG 和双万向节机构;此外,可以通过操纵剪刀式万向节来避开该系统中的外部奇点(饱和奇点)。该系统产生较小的扰动扭矩,并且除了原点和沿外万向节轴的线外几乎没有内部奇点。本文针对上述 CMG 系统提出了基于模型预测控制 (MPC) 的转向法则。所提出的基于MPC的转向律可以直接实时提供非线性约束下的最佳万向节速率,而不需要使用CMG雅可比矩阵的逆,因此,在确定该系统的万向节转向速率时不必考虑奇异性。转向律设计中考虑了三种目标函数:快速机动加权型、最小控制能量型以及这两种类型的混合型。通过数值模拟比较了这三种基于 MPC 的转向律的控制性能。结果表明,混合型既能实现控制初始阶段角速度上升时间短,又能平滑稳定目标姿态;而且其稳定时间和控制效果介于其他两种类型之间。
To overcome the singularity problem of control moment gyros (CMGs) and simultaneously provide a lightweight, agile, and simple dynamical system, a CMG system called a double-gimbal scissored-pair control moment gyro (DGSPCMG) was recently proposed. The DGSPCMG is a hybrid mechanism that combines scissored-pair CMGs and a double-gimbal mechanism; further, the outer singularities in this system (saturation singularities) can be escaped by steering the scissored-pair gimbals. This system generates less perturbation torque and has almost no inner singularities except at the origin and the line along the outer gimbal axis. This paper proposes model predictive control (MPC) based steering laws for the mentioned CMG system. The proposed MPC based steering laws can directly provide the optimal gimbal rate under the nonlinear constraints in real time without using the inverse of the CMG Jacobian matrix, and hence, the singularities do not have to be considered when determining the gimbal steering rate of this system. Three objective functions are considered in the steering law design: fast maneuver-weighted type, minimum control energy type, and a hybrid of these two types. The control performances of these three MPC-based steering laws are compared via numerical simulations. Results demonstrate that the hybrid type can achieve both a short rise time of the angular velocity at the initial stage of control, and smooth stabilization to the target attitude; moreover, its settling time and control effort is between those of the other two types.