Currently, most body sensor network applications are built upon an existing OS (TinyOS). The application acts as a middle ware between the sensors and the operating environment (OE). Our goal is to research and produce an OE specifically designed for use with many different medical sensors. The main part of this will be built using pieces from nanoRK and RT-Link. We would like the OE to have the following features:
Adaptive node scheduling: This would dynamically configure the sampling and sending rates for the sensors based on the situation. For example if a heart monitor detects a heart attack, the node would change the sampling rate to be more frequent and of a higher priority in order to get the most information possible.
Networking and data aggregation: If a sensor only needs to send 5 bytes at a time, and the information is not mission critical, aggregate the information into larger chunks to be send using fewer packets.
We'll begin by working on IMEC nodes, which provide a platform for us to test our software on. For an end goal, we would like to have the software running on the node and being used to dynamically change the sampling rate of a sensor attached to it. I will be working with Rahul Mangharam.