A Comparative Case Study of EMG-Driven Controllers in Transtibial Prostheses

Lower limb amputation greatly affects quality of life by restricting functional mobility. Despite advancements in prosthetic design, powered transtibial prostheses still have limitations in user control and adaptability to dynamic environments. This research presents a comparative analysis between a novel electromyography (EMG)-driven variable impedance controller (VIC) and a hybrid controller (HC) that integrates a volitional EMG-driven musculoskeletal model with a finite-state machine impedance controller. A Hill-type muscle model was used to model the gastrocnemius and tibialis anterior muscles. Biomechanical testing was conducted with a transtibial amputee to assess the controllers' performance across various tasks, including ambulation on level ground, stairs, and ramps, using EMG signals from the residual limb. Results demonstrated that the VIC provided more repeatable performance, perceived control, and power output. Notable effect sizes for peak power, observed in ramp ascent (Cohen's d $= -1.04$ ) and high-speed level ground walking (Cohen's d $= -2.92$ ), illustrate robust differences in joint-level output even when walking speeds and cadences were comparable. The greater predictability of the VIC led the user to feel more in control and comfortable throughout the various activities. On the other hand, the HC controller performed better in enabling more seamless transitions between gait subphases, particularly during stair ascent, which led to a significantly higher ROM ( $18.63~\pm ~1.53$ deg vs. $12.43~\pm ~1.86$ deg) and nearly double peak power compared to the VIC. This comparison lays the groundwork for future research into optimizing EMG-based control strategies that adapt to both biomechanical demands and user preferences.