Force impedance control method integrating feed-forward compensation and variable damping modeling

The invention discloses a force impedance control method fusing feed-forward compensation and variable damping modeling, under a position impedance control framework, adaptive feed-forward PI compensation and a deep neural network variable damping model are fused, and the adaptive PI control effectively compensates reference position information and reduces force steady state errors through feed-forward control; and the deep neural network variable damping model adjusts the damping coefficient on line in real time so as to dynamically adapt to an unknown contact environment and realize effective control of the contact force. 本发明公开一种前馈补偿与变阻尼建模相融合的力阻抗控制方法，在位置阻抗控制框架下，将自适应前馈PI补偿与深度神经网络变阻尼模型融合，其中自适应PI控制通过前馈控制，有效补偿参考位置信息，降低力稳态误差；深度神经网络变阻尼模型在线实时调整阻尼系数，以动态适应未知的接触环境，实现接触力的有效控制。