Motorized FES-induced cycling is a rehabilitation treatment for individuals with neurological impairments. An EMG sensor better integrates the human subject in the robotic system than in preceding cycling experiments. Signal processing techniques are utilized to remove noise from FES intensities. EMG tracking is utilized to provide individualized stimulation patterns to 4 muscles. An adaptive learning-based approach is explored to estimate the unknown muscle activation time constant. Cadence tracking ensures constant speed. A passivity-based analysis is developed to ensure stability in both EMG and cadence controls