A Knee Injury Prevention System by Continuous Knee Angle Recognition Using Stretch Sensors

The knee is the most common site for running injuries. Fatigue causes changes in the knee angle at landing, which increases the impact acceleration at landing and leads to injury. To prevent such injuries, this paper proposes a wearable system that acquires knee bends and estimates the knee angle in real-time during running using a supporter-shaped device equipped with a stretchable sensor. The result of the running experiment outdoors on flat and sloping roads shows that the proposed system can present the double-knee action, which is an indicator of the knee bending in running. For knee angle recognition based on treadmill running data, the average MAE was 5.8. in ranged from approximately 4 to 8. individuals, when using training data of the same subject's other trial. For recognition of the knee angle during running under fatigued condition, the result shows that the proposed system can recognize the knee angle with MAE approximately 5.0, when using training data of the user's own non-fatigued data.