Wireless medical sensors – context, robustness and safety

Wireless medical sensors are an emerging technology. Wireless sensors form networks and are placed in an unknown environment. For indoor scenarios context detection of medical sensors, e.g. removal of sensors from a specific room, is important. Current algorithms for context detection of wireless sensors are based on RF signals, but RF signal propagation and room location show only a weak correlation. Recent approaches with RSSI-measurements are based on prior fingerprinting and therefore costly. In our approach, we equip wireless sensor nodes with a barometric sensor to measure pressure disturbances that occur, when doors of rooms are opened or closed. By signal processing of these disturbances our proposed algorithm detects rooms and estimates distances without prior knowledge in an unknown environment. Based on these measurement we automatically build a topology graph representing the room context and distances for indoor environment in a model for buildings. We evaluate our algorithm within a wireless sensor network and show the performance of our solution.