Wearable sensor-based fuzzy decision-making model for improving the prediction of human activities in rehabilitation

Abstract

Sports actions are commonly recurrent due to the abnormal dynamic human activities. Detecting physical injuries based on the actions of the sportsperson helps to fasten rehabilitation treatments. Rehabilitation relies on the precise detection of activities and continuous monitoring of the actions of the sportsperson. In this paper, wearable sensor-based fuzzy decision-making (FDM) model is introduced for improving the prediction accuracy of different activities of the sportsperson. This model relies on altering sensor data aggregation and processing them using classification conditions for improving the prediction accuracy. The decision-making is performed by linearly classifying independent membership functions for different aggregation time and inputs. The combined processing of the inputs and time-based actions using independent decisions helps to improve the prediction accuracy of 93.3% with 26.081 ms decision time compared to conventional algorithms.