The availability of a new generation of accurate, low-cost sensors to scientists has resulted in widespread deployment of these sensors in a variety of environments. Data generated by these devices are often in a raw, proprietary or unstructured format. As a result, it is difficult for scientists to analyse or query across various biological and physiological sensor data values. There exists both a physical-digital divide between sensor data with related real-world conditions, and a knowledge divide between the information needs of domain specialists. A key challenge is to bridge these divisions in order to allow scientists to make better decisions based on the sensed in- formation. The goal of this research is to show that low level data collection resources such as sensors can be used for high level query expressions and knowledge extraction, without the need for expensive human based operations. To achieve this goal, it was necessary to deliver a generic approach to enriching raw sensor data, providing information services to enable the end user to acquire knowledge from low-level sensor data and defining an integration framework for sensor data and related contextual information. As a result, key research questions of interpreting heterogeneous sensor data, enriching sensor data with contextual information, and integrating sensor data with related participant and environmental information, are tackled over the course of this dissertation.