In this paper, an adaptive neural network command filtered backstepping impedance control method is developed for uncertain robotic manipulators with disturbance observer. First, an adaptive neural network algorithm is used to estimate the uncertain dynamics in the robot system. Second, impedance control is introduced to adjust the force and position relationship in physical human–robot interaction (pHRI). Third, a disturbance observer is employed to estimate the unknown external disturbance in the environment and compensate the control system to improve the safety of pHRI. Then, the command filtered technique can overcome problems of the ‘computational complexity’ and ‘singularity’ of traditional backstepping design. Finally, the simulation results are provided to illustrate the effectiveness of the proposed control method in pHRI.

本文针对具有扰动观测器的不确定机械手，提出了一种自适应神经网络指令滤波后推阻抗控制方法。首先，使用自适应神经网络算法来估计机器人系统中的不确定动力学。其次，引入阻抗控制来调整人机交互中的力和位置关系（pHRI）。第三，采用扰动观测器来估计环境中未知的外部扰动并补偿控制系统以提高pHRI的安全性。然后，命令过滤技术可以克服传统反推设计的“计算复杂性”和“奇异性”问题。最后，提供了仿真结果以说明所提出的pHRI控制方法的有效性。