Personalized Fitness Assistance Using Commodity WiFi

Building upon the insights gained from the exploration of activity monitoring using WiFi signals, this chapter now focuses on the development of personalized fitness assistance using WiFi signals specifically designed for home/office settings. Traditional approaches, which often rely on costly wearable sensors or specialized hardware installations, can be intrusive and uncomfortable for users. In light of these limitations, our research aims to provide a more affordable and user-friendly solution. To achieve this, in this chapter, we develop a personalized device-free fitness assistant system that utilizes existing WiFi infrastructure in home/office environments. Our system aims to provide personalized fitness assistance by differentiating individuals, automatically recording fine-grained workout statistics, and assessing workout dynamics. Using deep learning techniques, we perform individual identification based on workout interpretation, enabling tailored assistance. Our system also analyzes short and long-term workout quality and provides insightful workout reviews for users to enhance their exercises. We ensure system robustness through a spectrogram-based workout detection algorithm and a Cumulative Short Time Energy (CSTE)-based workout segmentation method. To assess the performance of our system, we conducted extensive experiments with a significant number of participants. The results reveal a high level of accuracy in both workout recognition and individual identification, underscoring the effectiveness of our system in providing personalized fitness assistance.