Design of a wearable multi-sensor system for physical activity assessment

In-situ measurement of physical activity of human under free-living environment is important for assessing their exposure to the environment. Innovative sensing technology is needed to realize such measurement. This paper presents a wearable, multi-sensor integrated measurement system (IMS) for assessing physical activity. Detailed analysis and simulation were performed for the parametric design of the IMS hardware and embedded software. An adaptive-scheduling method was devised to improve the battery power efficiency, reducing the energy consumption by over 16 times as compared to the conventional technique, enabling continued system operation of up to 54 hours. Experiments on human subjects have shown that the multi-sensor IMS is more effective in recognizing activities of varying intensity than the traditional activity measurement devices that use accelerometers alone. Error analysis of the IMS accelerometer outputs is also presented.