The main advantages of smart wearable devices are their convenience and real-time nature, making them a great potential for the quantitative management of daily sports activities. In this study, a sports pattern recognition model based on smart wearable devices is proposed, which aims to recognize and classify different sports activities by collecting accelerometer and gyroscope sensor data, combined with feature extraction and classification algorithms. First, pre-processing operations such as denoising and normalization are performed on the collected data, and time domain features such as variance and peak are used for feature extraction. Then, Particle Swarm Optimization Support Vector Machine (PSO-SVM) model is used for training and classification. The experimental results show that the PSO-SVM model has a significant advantage over the traditional GS-SVM model in action recognition. Specifically, the average recognition rate of 14 sports actions is 94.55%, and the recognition rate of each sport is more than 85%. In addition, the training time of PSO-SVM is also significantly shortened compared to GS-SVM. Based on these results, this paper demonstrates that the proposed model has higher accuracy and practicality in practical applications, especially in the quantitative management of daily sports activities. The findings provide strong support for the application of smart wearable devices in the field of health management.
