Project: Mobile Environmental Sensing System Across Grid Environments (MESSAGE)
Reference: SRT 7/5/7
Last update: 12/07/2012 14:06:07
To harness the potential of diverse, low cost and ubiquitous environmental sensors to provide data to address key scientific challenges in the field of transport and environmental monitoring and modelling and analysis.
To develop a flexible and reusable e-Science infrastructure to support a wide range of scientific, policy related and commercial uses and applications for the resultant data and to demonstrate the operation and utility of this infrastructure in diverse case study applications.
The project will gain a clearer understanding of how pollutants behave in urban areas so as to provide underpinning knowledge needed to put in place policies aimed at achieving the DfT objective of respecting the environment.
Overall, DfT will not impose any particular methodologies or break points. The project is jointly funded by EPSRC as part of their e-Science programme, with EPSRC contibuting a further £1.7m making the total cost of the project £3.4m.
The Project will be jointly funded within the DfT, between the Chief Scientific Adviser's Unit and Transport Technology Standards Division.
University of Southampton
Transportation Research Group, Dept of Civil and Environmental Engineering, Southampton, Hampshire, SO17 1BJ
Univ of Leeds, Leeds, LS2 9JT
Imperial College London
Exhibition Road, London , SW7 2AZ
020 7589 5111
University of Newcastle
6 Kensington Terrace, Newcastle upon Tyne, UK, NE1 7RU
University of Cambridge
The Old Schools, Trinity Lane, Cambridge, UK, CB2 1TN
Cost to the Department: £1,716,000.00
Actual start date: 13 February 2007
Actual completion date: 31 October 2009
MESSAGE Project Website
Author: Imperial College London, Newcastle University, University of Cambridge, University of Leeds and University of Southhampton
Publication date: 30/09/2008
Source: MESSAGE Project Website
More information: http://bioinf.ncl.ac.uk/message/?q=node/5
Summary of results
- Environmental data includes a variety of information about our surroundings such as temperature, noise, concentrations of pollutants and weather. This information says much about the areas in which we live and work and an extreme environment, whether good or bad, can have a significant impact on the individuals within it. This is particularly so in urban areas where there is an almost continuous movement of large numbers of individuals and vehicles. While there is limited scope for control of some environmental factors such as the weather, others such as noise and air pollution can be managed in order to offer improvements for local inhabitants.
MESSAGE is specifically targeting the detection and modelling of noise and air pollution data. High atmospheric concentrations of local air pollutants such as Carbon Monoxide (CO), Nitrogen Dioxide (NO2), Particulate Matter (PM) and Ozone (O3) have been shown to be harmful to health, both through epidemiological studies and laboratory exposure research (WHO, 2005). Due to their localised impact, the geographical distribution of these pollutants is important and acute exposure episodes tend to be more prevalent in urban areas where road traffic is a significant source of pollution (DEFRA, 2004). These impacts are substantial, indeed a recent report by the European Environment Agency (EEA, 2007) pointed out that in Europe the effect of premature deaths from air pollution-related illnesses now exceeds that from road accidents.
Urban pollutant concentrations are primarily a function of the local source emission rates from both mobile (e.g. vehicle tailpipe emissions) and static sources (e.g. local industries). Traditional assessment involves the long-term average modelling of these sources combined with relatively sparse networks of fixed monitors (such as the London Air Quality Monitoring network) with measurement protocols defined by the current air quality legislation (e.g. DEFRA, 2007). However, isolated studies have indicated that the pollutant concentrations within a street may vary by an order of magnitude over the space of a few meters and over the period of a few seconds (e.g. Dobre et. al. 2005). Moreover, pollution hotspots have been associated with traffic management interventions (e.g. signalized junctions) and alternative control strategies can be shown to have an impact on pedestrian exposure levels (Ishaque, 2007). A better understanding of the distribution of these episodes may on the one hand lead to further insight into the mechanisms by which air pollution affects health, while also aiding the development of effective strategies to mitigate these impacts.
Such studies require the integration of data describing pedestrian activity, traffic, weather and air pollution levels with far higher spatial and temporal resolutions than is typically available. If such data can be gathered and even processed in realtime, it becomes possible to make almost real-time decisions that can be applied to control of the urban environment and hence control the pollution within that environment. An example would be the dynamic changing of traffic signal timing to alter the flow of vehicles within an area in order to control the pollution emissions and perhaps allow the dispersal of a pollution hotspot.
Provision of such data using traditional, fixed-location, high accuracy monitoring systems is infeasible as such systems are bulky, expensive to purchase and expensive to maintain due to the need to regularly re-calibrate sensors. Simultaneous advances in communications, positioning, computing, sensing and modelling technologies have opened the possibility to develop a pervasive, mobile, wireless environmental sensing network and data processing infrastructure. In MESSAGE we are developing a heterogeneous sensor network that includes both mobile and static sensors. Static sensors provide readings from a guaranteed fixed location but cannot realistically be deployed in sufficient volume to cover a whole urban area at the necessary granularity. Mobile sensors move around constantly within the urban environment providing readings that 'fill in the gaps' between information provided by static sensors in order to offer much more environmental data without the volume of sensors that would be necessary in a static deployment. Further, by varying the types of sensors and deploying them in significant numbers, a complex data store can be built-up containing a large amount of information for future modelling. Moreover, the underlying architecture that is used is generic and may be applied in a variety of other areas where sensing and processing large quantities of data is necessary.