基于单个神经元的机器人自适应控制

Adaptive Contorl of Robotic Manipulator Using a Single Neuron

本文提出了针对多关节机器人采用单个神经元的独立关节控制方案，控制器由ＰＩＤ反馈和ＰＤＤ２前馈补偿构成，关节之间的静态和动态耦合通过单个神经元控制器前馈补偿．对一个二自由度机器人采用这种控制策略进行仿真．仿真结果表明该控制器在有非线性耦合、非线性摩擦、负载变化、模型参数变化时仍能跟踪期望轨迹．

This paper presents an independent joint control scheme for multi-joint robot manipulators using a single neuron. The controller consists if PID feedback plus PDD2 feedfoward. The static and dynamic couplings between the joint motions are compensated by the feedfoward of a single neuron. Simulations are presented to a two-DOF manipulator using the proposed method. The results confirm that trajectory tracking is achieved despite nonlinear coupling, nonlinear friction, load changing and parameter changing.