Feasibility analysis of a double leg perturbator robotic setup for the purposes of joint impedance estimation

Joint impedance is an important metric used for robotic gait trainers and rehabilitative therapies used to improve the walking ability of neurologically impaired people. Joint impedance is an important metric which describes the relation between resultant joint torques and changes in position, velocity and acceleration, thus influencing the movement and posture of the joint. Though the joint impedance of either the ankle, the hip, or the knee in the leg have been assessed during walking in isolation, thus far the joint impedance values of all three joints of the leg have not been succesfully identified simultaneously. Van der Kooij et al. identified the hip and knee joint impedance values accurately during the swing phase using the Lower Limb Perturbator (LOPER) [1]. However, this setup did not accurately estimate the ankle joint impedance. In this research the LOPER was expanded with an extra perturbator attached to the shank to assess the feasibility of using a double pusher setup to identify the joint impedance values of all joints in the leg during the swing phase of walking. Participants walked on a treadmill with the LOPER attached and were given 4 different types of perturbations using the double this expanded double pusher setup of the LOPER. The feasibility of this setup was tested by assessing two requirements of the setup: the transparency, and the joint angle response. The transparency was assessed using three criteria set in previous research. These were: the root mean square interaction force during unperturbed walking should not exceed 10 N, the peak interaction force during unperturbed walking should not exceed 20 N, and the joint angle deviations created by walking with the LOPER should remain smaller than the difference found between different participants when both walk without the LOPER. The joint angle responses were tested using a newly set criterion in this research based on the results of previous research. This criterion stated that the peak joint angle response of the perturbations should at least be 0.07 rad. All three conditions concerning the transparency were not met in testing due to problems with the stability of the LOPER setup during testing. The fourth criterion concerning the joint angle response was met for one perturbation condition. No joint impedance estimations have been done due to the transparency being insufficient. However, if a new controller would be implemented in the expanded LOPER setup, then the combined perturbations on the thigh and shank do show promise in sufficiently perturbing the leg for the purposes of joint impedance estimation.