Estimating Control Torque for Postural Balance Among Unilateral Lower Limb Amputees During Different Visual Environments

Lower limb impairment leads to the loss of sensory feedback and, as a result, poor postural abnormalities compared to able-bodied individuals. To maintain postural balance during gait initiation, individuals with lower extremity amputation over-depend on visual inputs and somatosensory information from the sound limb. This additional demand leads to higher cognitive load and additional extension and flexion torque in the healthy limb, leading to the unacceptability of current prostheses. In this study, a biomechanically relevant ankle-foot prosthesis model with a fractional order filter-based internal model control-based proportional integral derivative approach was explored to estimate the required control torque on the prosthetic ankle joint during dynamic weight-shifting activities in different visual environments. The results indicate the effectiveness of the proposed approach in adapting the balance uncertainties in different visual conditions. Pilot findings during weight shifting exercises show that the proposed
${f}$
-IMC-PID and traditional IMC-PID controllers present a maximum RMSE between the experimental and achieved lean angles of 0.01371° and 0.08414°, respectively. In addition, the present work compares the two baseline postural models: the shock absorption inverted pendulum model and the inverted pendulum-based virtual model. The present work could potentially reduce the rehabilitation duration by addressing the design-oriented challenges in prosthetic solutions.