受限机器人神经形态的变结构位置／力控制

Neuromorphic Position/Force Control for Constrained Robot Manipulator Using Sliding Mode

本文基于Ｊｅａｎ和Ｆｕ（１９９３）建立的受限机器人模型的降阶形式，利用变结构系统理论，设计了具有未知动态的受限机器人轨道／力追踪控制的一个神经形态位置／力控制器．提出的学习方法仅仅利用了机器人动态模型的一般结构，不需要其精确信息，计算迅速，易于实现，仿真结果验证了提出的方法的有效性．

A neuromorphic position/force controller using the theory of variable structure system is presented for tracking of trajectory/force control of constraint robot with unknown dynamics based on its reduced form presented by Jean and Fu (1993).The proposed learning scheme only makes use of a priori knowledge of the general structure of the robot dynamics, it does not require the exact knowledge of robot dynamics,and is computationally very fast and amenable to parallel processing implementation.A twolink robot manipulator with a circular path constraint verifies the good performance of proposed learning algorithim.