A Code Generator for Distributing Sensor Data Models

As wireless sensor networks mature, it becomes clear that the raw data collected by this technology can only be used in a meaningful way if it can be analyzed automatically. Describing the behavior of the data with a model, and then looking at the parameters of the model, or detecting differences between the model and the real data, is how experimental data is typically used in other fields. The work presented here aims at facilitating the use of sensor data models to describe the expected behavior of the sensor observations. The processing of such models can be pushed into the wireless sensor network to eliminate redundant information as early in the data collection chain as possible, thus minimizing both bandwidth requirements and energy consumption.