Soft Real Time Data Driven IoT for Knee Rehabilitation

Abstract

This article presents soft real time data driven Internet of Things (IoT) for knee rehabilitation using cyber physical sensory information system interfaced with cloud storage. Custom made wearable wireless motion capture suit interfaced to smart watch as the IoT are built for biofeedback visualization. Mullti-sensor integration and data fusion mechanisms are employed to obtain input vectors of knowledge base and the output vector is based on patient classification defined using multivariate statistics by the healthcare professionals. Case based reasoning is applied for the established reference standard in order to produce patient centric actual knee rehabilitation status and classification using semi supervised deep learning method. Wearable IoT is automatically updated the actual knee rehabilitation status and classification of a patient using relevant cyber physical sensory information retrieved from the cloud storage connected vis AWS cloud. Hence, a soft real time data drive IoT for knee rehabilitation system is successfully tested and validated using semi supervised deep learning cyber physical sensory information database subject to statistically quantified parameters by health professionals based on principle component analysis and patient centric parameters based on independent component analysis. The data driven IoT built has been validated in rehabilitation clinics by relevant physiotherapists and patients with the average age of ±36.8.