Automatic Real-Time Tracking of Soleus Muscle Fascicles for Ankle Exoskeleton Control

Abstract

Human locomotion, like walking and running, is a complex and energy-intensive task, and exoskeletons are being developed to reduce the metabolic cost of locomotion. Dr. Pridham and Dr. Stirling recently developed an ankle exoskeleton control algorithm that activates alongside energy usage in muscle cells, and it requires real-time tracking of soleus muscle fascicles to estimate energy usage. This project presents a computer vision algorithm that tracks muscle fascicles from B-mode ultrasound images. The algorithm is based on affine optical flow and operates at a framerate sufficient for real-time use. It also includes a heuristic to mitigate tracking drift, improving the robustness and accuracy of the tracking. The ability to track muscle fascicles in real-time is crucial for implementing the exoskeleton control algorithm in a real system, and this computer vision algorithm lays the groundwork for doing so. Long-term, this technology could have significant implications for individuals with mobility impairments or those who are always on their feet.