Research On Control Of Contact Forece Between Robot And Unknown Environment Based On Impedance Control

Robots are expected to participate in and learn from long term interaction with humans,and work with people to accomplish the required tasks ranging from elderly care,entertainment to education.In all of these applications,robots which are stiff and tightly controlled in position will face problems such as saturation,instability,and physical failure,when they interact with unknown environments.Impedance control is considered as an effective method to control the interaction between the robot and the unknown environment,but the key problem is that the dynamics of the robot is usually difficult to model in the design of impedance control.In order to solve this problem,the iterative learning and neural network were applied to impedance control.Impedance control design based on both iterative learning control and neural network,the impedance control method is applied to a human-robot cooperation situation which verify the effectiveness of the proposed algorithm.The main research contents and innovative results are as follows:1.Introduce an error signal between the real system and a virtual system with specified impedance model.From that,when the error signal becomes zero,which indicate that the dynamics of the robot arm is governed to follow a target impedance model and the control objective is achieved.Furthermore,a variable z containing the impedance error is defined with which the control objective becomes more compact.The following control design and analysis will show that z in impedance control is ‘equivalent' to the position error in position control.Therefore,it becomes possible to extend some existing methods in position control to impedance control.The definition of the impedance error makes the control target more explicit and provides convenience for subsequent derivation.2.Combine the iterative learning with impedance control and design the impedance control method based on iterative learning.First,the learning impedance control based on Linear-In-parameters(LIP)property is designed,and the convergence of the impedance error is verified by a rigorous proof.Further,the robot dynamics is mainly compensated by the high gain scheme.Finally,the difference between the impedance control based on the LIP and the high gain method is briefly contrasted,and the selection method is given.3.Neural Networks based impedance control is proposed.At first,Neural Networks is employed to approximate unknown and uncertain robot dynamics system,so as to realize the model free impedance design.The convergence of the impedance error is verified by a rigorous proof.The performance and stability of the system are discussed and verified by simulations.4.A human robot collaboration scenario is established to estimate the human motion intention based on neural network.The estimated motion intention is integrated into the adaptive impedance control,which enables the robot to track a given target impedance model,which enables the robot to actively follow the human motion.