Given their increasing level of complexity, today’s modern phones combine the functionality of countless, originally standalone devices. However, the measurement of human physiological parameters makes an exception as it requires locally placed sensors. Measurement devices as such – or as they usually called, ’wearables’ - should be as small as possible not to be uncomfortable. In addition to being small in size, long operational time is also a usual requirement. This can be achieved by low power consumption, therefore functionality should be minimized. Wearables usually only measure and transmit data but let the processing be performed by a phone nearby. Performing pre-processing on the wearable might only take place if it reduces power consumption by decreasing the amount of data to be transferred.
As a collaborative work with a mobile software services company, this thesis aims to present an easily deployable and customizable sensor system that can be used for various sensor based mobile applications.
The development of a universal, expandable and customizable system will be shown. The system builds on existing mobile and radio communication technologies with special attention to low energy protocols. The development will be presented in bottom-up approach: starting with hardware, following with client application development getting to a high level signal processing framework as an end-to-end demonstration of multi-level interoperability. A demo application will also be presented which was designed to demonstrate a comprehensive overview of functionality. The demo focuses on motion analysis, performing numeric gesture recognition using machine learning based classification, forwarding results to a web server or a graphical interpreter. The two major features of the complete system are low energy operation and small size: the assembled circuitry can even fit into a wristband or a ring. The system described in this thesis is subject to continuous development, which might result in a future market ready product.