Control of a Self-adjusting Lower Limb Exoskeleton for Knee Assistance

In this paper, the authors present the control architecture of a multi-contacts lower limb exoskeleton which is designed for knee rehabilitation purposes. The device's novel mechanical architecture, which comprises passive mechanical linkages connecting the user limbs to an external rigid mechanical structure, allows amore effective control of the system's transparency as only principal torque components are transmitted to the user's anatomical joints. Different types of sensors were used to capture the user's kinematics as well as to detect contacts between the user's soles and the ground. This data allow the estimation of the gait phases in realtime, using Principal Components Analysis (PCA) and Markov chain. Finally, a predictive controller was implemented on the device to assist the knee joints during specific gait phases.