For individuals with neuromuscular disorders (NDs) affecting the coordination and control of their legs, motorized functional electrical stimulation (FES) cycling serves as a rehabilitation strategy and offers numerous health benefits. A motorized FES cycle is an example of a hybrid exoskeleton involving cooperative physical human–robot interaction where both the cycle’s motor and rider’s muscles (through electrical stimulation) must be controlled to achieve desirable performance. A robust sliding-mode cadence controller is developed for the rider’s muscles, while an adaptive admittance controller is employed on the cycle’s motor to preserve rider comfort and safety. A switched systems stability analysis using Lyapunov and passivity-based methods is conducted to ensure global asymptotic stability of the admittance error system and that the cadence error system remains passive. Experiments are conducted on one able-bodied participant and four participants with NDs to illustrate the performance of the control design. For the participants with NDs, the controller achieved an average admittance tracking error of −0.08±1.05 rpm at an average cadence of 47.85±1.13 rpm for the desired cadence of 50 rpm.