Movement Balance Evaluation for Basketball Training Through Multi-Source Sensors

Abstract

Balance ability is the basic sports quality of athletes. For basketball players, balance training includes take-off, turning, confrontation, shooting, landing, and other links. If the players have good balance ability, they can effectively prevent sports injury and competition interference and improve the performance of basketball competition. This paper adopts the acceleration signals from multi-source sensors to evaluate movement balance for basketball training. First, acceleration signals are collected by acceleration sensors to depict the basketball player's actions. Second, the hidden Markov model is used to describe the change or transfer of different states during player's actions. Third, the acceleration signal and observation sequence from hidden Markov are used to determine whether the player is under imbalance state. The effectiveness is evaluated on a private dataset.