Design and implementation of a countermovement jump performance estimation system using a wearable device with IMUs based on machine learning algorithms

The rapid evolution of technology has led to a proliferation of wearable devices. These devices have a wide range of applications in hospital health care, health monitoring, and sports. In sports science, it is possible to assess an athlete’s neuromuscular function via explosive power in the legs, and the most common test for this is the counter-movement jump (CMJ). Traditionally, to assess the explosive power of an athlete’s legs, researchers ask the athlete to stand on a force plate and perform a CMJ several times. From this, seven variables are calculated using the force-time curve measured by the force plate, and then these variables are interpreted by sports professionals to evaluate the goodness of the CMJ. However, a force plate is expensive and difficult to transport. Taking this into account, this paper designs a machine learning-based model to predict CMJ performance variables via an economical and wearable device with inertial measurement units (IMUs). The experimental results demonstrate that the proposed method is capable of predicting most performance variables of CMJ with an acceptable rate of error. Finally, we also implement a real-time performance variable measurement system to demonstrate the applicability of the proposed method.