Admittance Control of a Robotic Exoskeleton for Physical Human Robot Interaction

In this paper, an admittance control scheme is proposed for physical human-robot interaction with human subject's intention motion as well as dynamic uncertainties of the robotic exoskeleton. Human subject's intention motion is represented by the reference trajectory. Online estimation of the stiffness is employed to deal with the variable impedance property of the exoskeleton manipulator. An admittance control approach is firstly presented in order to generate a differentiable reference trajectory in interaction tasks. Then a stability criterion can be obtained by the proposed control method. The designed controller includes linearly parameterization and estimation for the unknown items of the dynamics. Bounded and convergent errors are shown in the motion tracking. The experimental results indicate that the proposed control approach enables the human subjects to execute an admittance control task on the exoskeleton robot effectively.