Gait Analysis Using Virtual Body Sensor Networks for Biomedical Applications

In BSNs Signal processing usually consists of multiple levels of data abstraction, with raw sensor data to data which is calculated from each processing steps that includes feature extraction and classification. Here we present a multi-layer task model based on the concept of Virtual Sensors in order to improve design reusability and architecture modularity. Virtual sensor networks (VSNs) is an emerging form of collaborative wireless sensor networks. Virtual Sensors are abstractions of components of BSN systems that involve sensor sampling and tasks processing and issues data upon external requests. The model of virtual sensor implementation depends on SPINE2, which is an open source domain-specific framework that is developed to support distributed sensing operations and processing of signal for wireless sensor networks and enables code efficiency, reusability and application interoperability. This proposed model is applied in the framework of gait analysis through wearable sensors. According to SPINE2basedVirtual Sensor architecture a gait analysis system is developed and it is experimentally evaluated. The results obtained confirm that great value can be achieved to design and implement BSN applications through the Virtual Sensor approach at the same time maintaining high efficiency and accuracy.