Multi‐sensors edge computing for tennis stroke recognition in IoT environment

This paper proposes a multi‐sensors edge computing approach for tennis stroke recognition in an IoT environment. By applying the edge computing capabilities to this framework, we can ensure swift local processing of sensor data, thereby significantly reducing latency and enhancing system scalability for tennis stroke. In this way, we develop an advanced multi‐sensor based tennis stroke recognition mechanism, which employs dedicated measuring sensors mounted on the player's forearm to comprehensively capture intricate motion patterns and send the collected data to the edge computing center. Through rigorous analysis at the edge, we can determine the optimal actions in stroke detection with minimum delay, in which the optimal training datasets for stroke classification are also determined. Experimental results indicate that the proposed mechanism surpasses the state‐of‐the‐art methods and achieves effective adaptability and scalability for tennis stroke recognition in IoT‐driven environments, evidenced by their commendable performance even when applied to athletes not included in the training datasets.