Designing a Wearable Sensor System for ACL Injury Detection and Prevention in Alpine Skiing

Alpine skiing, a popular recreational activity enjoyed by millions worldwide, has witnessed substantial reduction in tibial fractures in recent decades owing to advancements in equipment technology. Despite safety improvements to protect the tibia, there has been a 280% increase in anterior cruciate ligament (ACL) injuries. Existing research shows that successfully releasing the boot from a ski relieves force on the skier’s legs and saves them from ACL injury. However, current bindings cannot detect loading on the ACL in order to reduce it, thus making skiers more susceptible to knee injuries in exchange for protecting the tibia.
Our team has developed a wearable knee sensor system and a release algorithm solution to reduce the frequency of ACL injuries. Our design features a dual-sensor system using IMU sensors and fiber optic cables to detect and record skier motion data predictive of an ACL injury and a rule-based decision algorithm that can trigger a release signal to ski bindings in real-time. We use four IMUs, two on each leg, to measure relative rotation between the tibia and femur. The IMUs are mounted using a small, hermetically sealed enclosure, strapped around the thighs and shins with velcro. Four fiber optic cables, which determine knee flexion, are sewn into leggings, looping around the knees. The sensor systems and decision algorithm are integrated with an ESP32 microcontroller and powered by a 3.7V battery, which are enclosed in a hermetically sealed box which should be robust enough to withstand strong impacts, low temperature, and other common hazards in a skiing environment. Sensor data is saved to a removable 8GB microSD card where it will be viewable as an animated 3D model in Blender. When an imminent ACL injury is detected, the release algorithm sends a signal through a 2-pin connector to the ski bindings. The hermetically sealed box which encloses the microcontroller and battery is designed to be robust enough to withstand strong impacts, low temperature, and other common hazards in a skiing environment through design features such as a circular gasket and O-rings.
Upon refactoring the geometry of the fiber optics cables, we were able to increase our resilience against flexural deformation, as required to withstand the strong vibrations of skiing. We were able to characterize the sensor’s nearly-linear performance as a function between voltage and angle. We achieved a minimal IMU lag of -0.88 degrees to characterize its inherent accuracy. Through experimenting with multiple velcro harnesses, we found a balance between comfort and secure IMU mounting with a 0.12° IMU/Leg per 1° Pants/Leg rotation. We also time tested the IMU subsystem and found an average signal delay of 32.8ms which is below the required 40ms maximum. Further tests of the system’s resilience to the skiing environments found that the battery lasts 5.5h at -15.1 °C which can be improved. Our tests concluded that, with the exception of the system battery life in cold weather, the accuracy of the IMUs and fiber optics, velcro harness mount and hermetically sealed box, and release signal delay, are capable in allowing the user to wear the system, accurately measure and record ski knee motion data, and trigger binding release when potential injury is determined. Future steps include full system testing with the Stanford Ski team, skier profile personalization in decision algorithm, and compatibility with common ski boots.