Experimental validation of a finite state controlled functional electrical stimulation walking system with real-time gait phase detection using a single wearable IMU sensor

Abstract

Functional electrical stimulation (FES) is an effective tool for activating lower-limb muscles in gait rehabilitation. Precise timing of FES according to gait sub-phases is crucial for effective gait movements. However, most FES systems are controlled by open-loop muscle stimulation based on pre-determined timings or therapist input. Real-time detection of gait sub-phases and controlling FES systems accordingly could improve efficacy. Yet many prior approaches either require multiple sensors or detect only two main gait events: Heel Contact and Toe Off. Simplification is essential for clinical translation, and a single-sensor setup can potentially streamline FES control in practical rehabilitation contexts. To overcome this limitation, the present study present a novel real-time gait phase detection algorithm for Finite State Machine (FSM) control during walking, utilizing a single wearable Inertial Measurement Unit (IMU). The algorithm was validated by recording stimulation timings and comparing them to lower-limb kinematics from simultaneous optical motion capture. Our algorithm accurately identified four gait sub-phases in real-time, with only a small differences relative to off-line gait sub-phase timings (averaging −2.88 ms for T1, 67.2 ms for T2, −0.68 ms for T3, and 6.63 ms for T4). We observed that most FES onsets occurred just after the gait phase transition, typically within 50 ms. Overall, this approach is simple to implement and shows potential for real-time FSM control of FES in gait rehabilitation, though additional validation is required before clinical deployment.