Presenter

Mr. Federico Dominguez - ETRO [Email]

Abstract

Modern sensor networks consist of spatially distributed sensor nodes that sense the surrounding environment and organize in a network. For environmental applications, they serve as a data gathering network that amplify and enhance our awareness and knowledge of our changing environment with minimum impact to it.
Environmental sensor networks have existed for decades and, despite this, several technological and scientific challenges still remain that hamper their widespread use. These challenges range from the design of efficient and effective autonomous nodes and communication protocols to the storage and distribution of the huge amount of data collected by these networks.
These challenges are even more pronounced in the noise pollution monitoring application domain where four dimensional real time data is desired with special emphasis on an accurate determination of the exact nature and location of the noise pollution sources.
This work contributes to the improvement of noise pollution monitoring sensor networks along three axes:
The development of low-cost self-testing noise sensors using off-the-shelf components. Low cost sensors allow for higher density network deployments and their embedded self-testing capacity compensate for their lower data quality and life cycle.
The development of a data fusion technique to locate noise pollution sources and generate accurate sound maps using a Wireless Sensor Network (WSN) equipped with microphone array sensors. A microphone array sensor is capable of measuring the noise pollution level (how loud) and directionality (where is it coming from) and when distributed over an urban area using a WSN it can generate an enriched sound map that identifies the location of noise pollution sources.
The development and prototype deployment of a software encapsulation layer that handles the distribution, visualization, and protection of the data generated by a noise pollution network. Using enterprise systems integration design patterns, the data generated by a typical environmental network (in the order of terabytes) can be securely made available to the public and the scientific community using intuitive geographical data visualizations and standard M2M communication protocols.
The end result is an architecture that allows us to build and deploy a low cost, high density noise pollution sensor network with the capacity to accurately locate pollution sources and effectively distribute this information to the public and stakeholders.

Short CV

Master of Applied Sciences and Engineering, Applied Computer Science, Vrije Universiteit Brussel, 2009