Signal Processing for Sensing
We are currently entering a new era of distributed communication, where distributed processing, communicating, and sensing are replacing more traditional, centralized architectures. The most important example of this revolution is appearing in the form of sensor networks: densely distributed networks of embedded sensors, controllers, and processors. This decentralized setting creates a plethora of new problems in signal processing and communications:
- Since communication is generally multipoint-to-multipoint, theseparation principle cannot be used in general.
- There is a clear interaction between the physics of the process that is sensed and the communication network.
- In the context of sampling, since spatial sampling kernels cannot be applied before sampling, there is always aliasing.
- Different tasks can be studied, namely, detection, estimation, classification, and control.
Our emphasis lies on the following topics:
- Distributed source coding schemes and application to hearing aids.
- Tomographic methods for estimating temperature and wind fields.
- Distributed sampling of physical phenomena, such as acoustic and electromagnetic fields.
- Analysis and modeling of time-varying channels.
- Design of low-power, low-complexity sensor networks, interfaces, and software.
- Outdoor sensor networks with application of sensor networks to environmental monitoring.
Within sensor networks for environmental monitoring we are currently contributing to the following projects:
- Collaboration with Prof. Marc Parlange’s Laboratory of Environmental Fluid Mechanics and Hydrology.
- The Nanotera project Opensense targeting mobile sensing for air quality monitoring (sensors on public transportation buses).
- Cooperative Visual Monitoring in Energy-constrained WSNs
- Efficient Data-collection in WSNs
- Opensense: Open Sensor Networks for Air Quality Monitoring
- Info4Dourou: WSN in Developing Countries
Past research projects:Back to research topics ]