Characteristic extraction of Tai Chi movement data—Based on self-powered wearable sensors

Although visual recognition has good recognition accuracy, it brings great hidden danger of privacy leakage. Although signal recognition has the advantages of device-free and privacy protection, it is sensitive to environmental noise and is not suitable for crowded environment, so sensor-based human behavior recognition is a more feasible choice. Therefore, this paper proposes a multi-level decision behavior recognition method based on self-powered wearable sensor fusion. In this paper, we propose a CM-WOA-based automatic dynamic sensor deployment optimization method for the feature extraction of Tai Chi action data. In behavior recognition based on wearable sensors, different deployment schemes of self-powered wearable sensors, will lead to different recognition accuracy, However, the traditional empirical deployment scheme cannot guarantee the best sensor layout. In order to further improve the recognition accuracy. In this paper, we propose a CM-WOA-based autodynamic sensor deployment optimization method for the feature extraction of Tai Chi action data, so as to find a balance between recognition accuracy and sensor deployment cost, and deploy as few sensors as possible on the premise of maximizing recognition accuracy. Finally, by comparing the scheme proposed in this paper with the other seven schemes, The feature extraction and recognition rate of Taijiquan movement data based on self-powered wearable sensor can reach 94%, which proves that the proposed multi-sensor deployment optimization method based on CM-WOA is effective in improving the overall recognition rate of the recognition model.