Elliptical trainers are prescribed in rehabilitative exercise but difficult to implement in populations with significant functional gait deficits. Typical elliptical machines do not mimic normal gait and therefore require modifications for clinical rehabilitation. This research builds on previous modifications of an elliptical trainer designed to simulate level-surface walking. This design differed from a commercial version. It included articulated footplates and an electromechanically-driven virtual-cam to control footplate position. Ankle dorsiflexion elicited lower-extremity muscle spasticity which produced an unwanted gait variant during stroke patient testing. Spasticity is a hyperexcitable stretch reflex causing inefficient gait. This project’s purpose was to develop an autonomous cam-profile adjustment to optimize the device’s rehabilitation potential. Foot-to-footplate forces were measured in stroke patients and compared to normative data. Greater than normal forces were considered spastic. An embedded controller was designed to reduce footplate forces via real-time cam-profile attenuation. A simulated spastic dorsiflexion load successfully proved the algorithm’s efficacy.