Energy-efficient algorithms for air quality monitoring using sensor networks
Project outside the universityHolst Centre (Eindhoven)
Project descriptionThe increased trend in global-urbanization has led to congested and polluted neighborhoods in our megacities. The major contributors to the deteriorating air quality in these urban areas include emissions from industries, transportation, home heating and garbage incineration. The combustion of hydrocarbon fuels in automobiles and airplanes produces toxic pollutants (such as nitrogen oxides, carbon monoxide), volatile organic compounds (e.g., benzene, toulene) and also large quantities of particulates, such as lead. An increased exposure to these pollutants causes respiratory diseases, eye infections, heart diseases and prolonged exposure could even lead to premature death. In addition, air-pollution is also related to increase in acid rains, global warming and ozone layer depletion. Thus, air quality monitoring is one of the key challenges for both healthcare and environment.
Traditional air quality monitoring solutions based on chemical analyzers, which although very accurate, are expensive and bulky. This not only hampers portability and deployment of these equipment to desired environments, but additionally offer poor spatial resolution. However, recent advances in technology has enabled the growth of cheap and affordable sensors and Wireless Sensor Networks (WSN), capable of sensing, calibrating and communicating the preprocessed data to a centralized source for further processing. A WSN for air quality monitoring would contain multiple nodes, with a diverse portfolio of sensors on-board. Each node in the network is in itself a heterogeneous platform containing sensors for air pollutants (such as CO2, NO2) and environment monitoring (such as Temperature, Relative Humidity, ambient light, sound). Such a network would offer spatiotemporal information on the air-quality of the environment, which must be processed efficiently for desired feature extraction. A diverse portfolio of algorithms cater to the needs for such a WSN, depending on the requirements of the application. For cold-start scenarios, these include Maximum-likelihood and Least squares based estimators. In case of long-term tracking, Artificial Neural Networks (ANN) and advanced Kalman filters could be employed. The proposed algorithm(s) must not only predict the air-quality and extract desired features of the environment but should also offer an energy efficient solution.
- Literature survey on the current state of the art for air quality monitoring
- Developing algorithm(s) for energy-efficient air quality monitoring
- Simulations to validate the algorithm
- Implementing the solution(s) on the IMEC IoT sensor network
- Testing and analyzing the performance of the implemented algorithm
- Thesis writing and documentation at IMEC-Holst Centre
- (Option) submit the work to a top-ranking publication
- Experience with signal processing/machine learning tools
- Proven experience with MATLAB/Python/C/C++
- Previous experience with wireless sensor networks and/or air quality monitoring is an added plus.
- Motivated student eager to work independently and expand knowledge in the field
- Good written and verbal English skills
LocationThe project is carried out at Imec Holst Centre, Eindhoven (contact: Raj Rajan).
prof.dr.ir. Alle-Jan van der Veen
Circuits and Systems Group
Department of Microelectronics
Last modified: 2016-01-22