Wearable Sensor-Based Exercise Monitoring System for Higher Education Students Using a Multi-Attribute Fuzzy Evaluation Model

Abstract

Exercise and Physical Activity are important factors to improve the student s health and academic status. Student exercise should be continuously monitored to eliminate risk factors and health issues. The previous monitoring system faced difficulties while handling the vast amount of data obtained from multiple sensors because it was affected by uncertainty and noise issues. The research difficulties are addressed with the help of the Multi-Attribute Fuzzy Evaluation Model (MAFEM), which monitors student s health using sensor data. The MAFEM approach uses the fuzzy set and fuzzy logic to derive the relationship between the features. In addition, the method uses preprocessing, fuzzification, defuzzification and rule evaluation processes. These steps are adjusted according to the threshold value that maximizes the personalization and holistic assessment efficiency because the system uses multiple attributes. During the analysis, MM-Fit dataset information is utilized to evaluate the system efficiency in which the system ensures the minimum computation complexity O(r.m.n) and minimum latency value (≈ 70mAh). In addition, the accuracy metrics are also applied to evaluate the system s effectiveness, with 97.11% precision, 0.23 RMSE and 0.26 MSE values.