空间非合作目标惯性参数在轨辨识

Inertia parameter identification of on-orbit uncooperative object

在空间冗余机械臂抓捕非合作目标的惯性参数辨识问题中,已有方法大都基于系统动量已知的假设,且辨识过程未考虑基座姿态稳定。针对目标动量未知的问题,设计了具有增量形式的惯性参数分步辨识算法。首先基于线动量方程得到关于质量和质心位置的第一组估计方程,采用增量形式消除未知线动量更新估计方程。辨识结果收敛后根据估计参数计算线动量估计值,代入以转动惯量为未知参数的第二组估计方程中,利用其增量表达式完成对转动惯量的估计。辨识过程中的激励由自适应零反作用控制输入提供,算法在保证基座姿态不受干扰的同时还能对惯性参数精确辨识。仿真结果表明,在30s以内算法已收敛,误差收敛到零的同时,基座姿态角速率控制精度在10-3以下,说明算法收敛快,精度高,同时还能实现基座姿态稳定。

Most algorithms of inertia parameters identification assumed the system momentum known and few attention was paid to the attitude stabilization of base. An incremental form of the identification equation was given based on the conservation principle of momentum and a fractional step method was proposed. Firstly, estimation equations about the mass and position of mass center were obtained and the unknown constant linear momentum was eliminated in incremental formulation. The estimate of linear momentum was calculated with the estimated parameters and then substituted into estimation equations about the inertia tensors. The incremental equations were obtained again to eliminate the unknown constant angular momentum. An adaptive reactionless controller was used to persistently excite the robot arm, which guaranteed the convergence of the algorithm as well as the stabilization of base attitude. Simulations results show that inertial parameters are accurately identified and the angular velocities of the base are less than 10^-3, which can be viewed as nearly undisturbed.