讨论了载体位姿无控情况下,漂浮基柔性关节空间机器人的关节运动控制器设计问题。利用系统动量、动量矩守恒关系及拉格朗日法,建立了空间机器人的动力学模型。为实现系统关节运动控制目标,借助于神经网络函数逼近技术,提出了一种柔性关节空间机器人的自适应反演控制方案。所提控制方案无需预知系统各惯性参数的真实信息,且可避免对载体位置相关量进行实时地测量与反馈,因此较适于实际应用。数值仿真结果表明:上述控制方案可使系统各柔性关节的振动较小,并能够有效地控制空间机器人完成所期望的关节运动。 The design of articulation controller of floating flexible joint space robot is discussed in the condition that the attitude of the carrier is uncontrolled. The dynamics model of space robot is established by using the relationship between system momentum, momentum moment and Lagrangian. In order to achieve the goal of system joint motion control, an adaptive inversion control scheme of flexible joint space robot is proposed by means of neural network function approximation technique. The proposed control scheme does not need to predict the real information of each inertial parameter of the system and can avoid the real-time measurement and feedback on the carrier position correlation. Therefore, it is more suitable for practical applications. Numerical simulation results show that the above control scheme can make the flexible joints of the system less vibration, and can effectively control the space robot to complete the desired joint motion.