Neural networks impedance control of robots interacting with environments

In this paper, neural networks impedance control is proposed for robot-environment interaction. Iterative learning control is developed to make the robot dynamics follow a given target impedance model. To cope with the problem of unknown robot dynamics, neural networks are employed such that neither the robot structure nor the physical parameters are required for the control design. The stability and performance of the resulted closed-loop system are discussed through rigorous analysis and extensive remarks. The validity and feasibility of the proposed method are verified through simulation studies.