Comparison of Different Fuzzy Controllers for Powered Prosthetic Gait

This paper compares the fuzzy controllers trained with three different data-based methodologies for the control of the active knee joint of a powered transfemoral prosthesis. For each methodology a different inference system structure and number of generated rules was produced. The control is optimized using adaptive neuro-fuzzy optimization techniques and it is aimed to work at different speeds within normal gait.