가변 강성 길항관절을 위한 적응 제어 알고리즘

Abstract

In this thesis, we consider the problem of estimating nonlinear stiffness of flexible transmissions in single link robots with antagonistic variable stiffness actuation. Joint stiffness estimation is obtained using an adaptive control algorithm. For the joint stiffness estimation, we assume that all rigid body dynamic parameters of robot except stiffness are known value. The motor position, velocity, link angle position, velocity and flexibility torque are assumed to be measurable for the state-feedback. An adaptive control algorithm with input-output linearization state feedback is used in our problem on the basis that this algorithm is optimal for our problem. Joint stiffness value is assumed to be intrinsically a nonlinear polynomial function of the deformation. Simulation results from performed of single link arm robots are reported, showing a good performance in trajectory tracking of link angle position and in estimating a nonlinear polynomial function of the joint stiffness.