Research On Adaptive Impedance Control Methods For Five-Bar Parallel Robot

Parallel robots have significant applications in industries and services due to their advantages of high stiffness,compact structure,large bearing capacity and good dynamic response.When a robot performs tasks such as grinding,handling,assembling,rehabilitation and surgery,a robot will come into direct contact with the environment(some person or operated object).But excessive contact force is prone to damage the robot or bring harm to the environment.Hence,a robot is required to have certain flexibility in the contact operation,which raises higher requirements on the robot control system.The impedance control is intended to establish a spring-damping dynamic system between expected trajectory tracking error and human-machine interaction,which is an important active flexibility control method.So it is of great value to design effective impedance control for parallel robot.However,the uncertainty and disturbance in the robot dynamics model affect the convergence of impedance error,and further affect the stability and robustness of impedance control.Therefore,it is vitally important to study the adaptive impedance control of parallel robot so as to improve the stability and robustness of impedance control for the purpose of enhancing flexibility,safety and reliability of the interaction between parallel robot and environment.The main study work and contributions of this paper are summarized as follows:(1)This paper proposes an adaptive neural network impedance controller with dead-zone modification.First,an impedance trajectory was established on the basis of the desired first-order impedance dynamic model,thus the design issue of impedance control was converted into a tracking control design issue.Then,based on the Euler-Lagrangian model of the five-bar parallel robot,an impedance control composed of a neural control term and a force control term was designed to ensure the convergence of impedance errors to zero or a small neighborhood of zero,which guarantees the achievement of the desired first-order impedance dynamics.Finally,Theoretical analysis and simulation results validate the effectiveness of the proposed adaptive neural network controller.(2)This paper proposes a composite learning impedance control scheme for robots with parameter uncertainties based on reference trajectory generation.First,based on a generated reference trajectory,the impedance control problem is converted into a special tracking problem.Then,based on the Euler-Lagrangian model of robot,an impedance control strategy composed of a composite learning control term and a force control term was designed to ensure the convergence of impedance error and the realization expected impedance dynamics.Finally,theeffectiveness of the proposed composite learning impedance control law is validated by theoretical analysis and simulations.(3)This paper proposes a disturbance-observer-based impedance controller.First,through a constructed reference trajectory,we convert the impedance control design problem into a particular trajectory tracking problem.Then,an impedance control strategy composed of disturbance observer,a PD control term and a force control term was designed to ensure the convergence of impedance errors to zero or a small neighborhood of zero,which guarantees the achievement of the desired second-order impedance dynamics.Finally,we verify stability of the proposed impedance controller by Lyapunov Theory and illustrate its effectiveness by simulation results.