On Detecting Freezing-of-Gait Through Sensor Data Fusion Using Evidence Theory

Freezing of Gait (FoG) is a common impairment that patients suffering from Parkinson disease have to experience in which the arms and legs of the patient freezes, rendering them incapable to resume walk that often leads to abrupt falls. Therefore, proper monitoring and detection of FoG events is of utmost importance to mitigate the undesired affect of such a disease. This work demonstrates the potential of using data fusion algorithm for timely detection of FoG and subsequent alert to the concerned authorities without delay through the use of Internet of Things (IoT) architecture. The architecture is composed of a smartphone, that is well suited and can be seamlessly integrated in society. The smartphone monitors gait parameters and alerts the medical authorities in case of emergency. Our proposed data fusion algorithm is based on evidence theory which requires very little storage and at the same time has very low execution time that are critical issues in real time IoT monitoring scenarios. We compared our results with four data fusion and classification models in the literature and observed that our model clearly outperforms the others producing 97.50% specificity, 99.87% precision and requires only 11 secs to run 28801 test samples.