Musculoskeletal Load Estimation
This application area focuses on estimating internal joint and musculoskeletal loads (e.g., shoulder and knee moments) from wearable sensors and contextual data. Instead of relying on laboratory-based motion capture systems and force plates, models infer the mechanical loads acting on joints during sports and daily activities using signals from IMUs, pressure sensors, and other wearables, often combined with basic anthropometric or subject-specific information. It matters because joint overuse and impact-related injuries are a major problem in both elite and recreational sports, as well as in populations with mobility impairments. Continuous, field-based load estimation enables individualized training prescription, early detection of harmful loading patterns, and more precise rehabilitation progression, all at scale and at lower cost than lab testing. Organizations use AI models to turn raw wearable data into actionable biomechanical insights that can be used by coaches, clinicians, and athletes in real time or near real time.
The Problem
“Estimate knee/shoulder joint loads from wearables—outside the biomechanics lab”
Organizations face these key challenges:
Overuse injuries appear “suddenly” because load is only tracked via rough proxies (minutes, distance, RPE)
Lab motion-capture sessions are expensive, infrequent, and don’t represent game-speed movements
Coaches get lots of sensor data (IMUs/GPS) but little biomechanical insight they can act on