| Popis: |
Aiming at the difficulty of force control in impedance control lies in the unknown parameters of the environment, a reference trajectory generation approach for impedance control is presented, which uses neural networks to estimate the environment parameters. In this method, the environment parameters are estimated by a neural network (NN1), which constructs the relationship between the environment deformation and its first and second derivatives, and the interaction force. Then, another network (NN2) constructs the relationship between interaction force and the deformation. The major advantage of the proposed method is that no exact environment model is required, so that it suits for operations on any unstructured environments. Finally, the method was tested in the pneumatic hydraulic combination compliant control system, which validated the validity of the proposed method. |
