MESSAGE: A Mobile Environmental Sensing System Across a Grid Environment Jeremy Cohen Department p of Computing p g Imperial College London OGSA-DAI Data Integration Meeting NeSC, Edinburgh 16th March 2009 The MESSAGE Project j • Mobile Environmental Sensing System Across a Grid Environment • 3 year project starting October 2006 • Funded jointly by EPSRC and DfT (~£4m), under EPSRC’s e-Science demonstration programme • 5 Universities, 19 industrial partners • Pioneering combination and extension of leading edge grid, sensor, communication and positioning technologies • Create radically new sensing infrastructure based on combination of ad-hoc mobile and fixed sensors • www.message-project.org MESSAGE Industrial Partners Equipment providers Service providers End users MESSAGE Objectives j To extend existing eScience, sensor, communication and modelling technologies to enable the integration of data from heterogeneous fixed and mobile environmental sensors grids in real time to provide dynamic estimates of pollutant and hazard concentrations. To demonstrate how these data can be usefully correlated with a wide range of other complementary p y dynamic y data on,, for example, p , weather conditions,, transport network performance, vehicle mix and performance, driver behaviour, travel demand, pollutant exposure and health outcomes. To implement relevant eScience tool sets and (fixed and mobile) sensor and communication system in a number of selected real-world case study applications, involving close collaboration with business and the public sector, and to thereby to demonstrate their value to the research and policy community. Technical Scenarios – Pollution Data Traffic Management Examining how increased environmental data can be used for better network control Key issues to be addressed... y Inclusion of environmental objectives j in Traffic Control y Resolution - reliability – cost tradeoffs y Distributed processing of air quality projections Traveller Information Disseminating g environmental data directly y to individual travellers Key issues to be addressed... y Data aggregation across wide spatial areas y Visualisation of temporally variant data y Interfacing systems with legacy applications Technical Scenarios – Pollution Data Health and Exposure Monitoring and reporting potential or actual exposure of individuals to pollutants Key issues to be addressed... y High spatial resolution monitoring y Distributed data buffering and processing y Path based data aggregation Pollution Modelling Using finer environmental data as inputs to increasingly detailed pollution models Key issues to be addressed... y Real Real-time, time, archive and predicted data analysis y Resolution - reliability – cost tradeoffs y Dynamic sensor resolutions e-Science Platform – Data Capture Data Capture Platform Reliable, efficient capture of data from an environment with an unreliable communications infrastructure and varying load load. y Different types of sensors, different pre-processing requirements y Different communications technologies y Real R l titime streaming t i and d iintermittent t itt t b burstt Scalable Cloud-based processing infrastructure Multiple sensor and communications technologies. Multiple DBs distributed across several sites. sites MESSAGE – Field Units 3 field unit types (Sensor + Pos + Comms) Newcastle • Electrochemical + GPS + Zigbee Cambridge • Electrochemical + RSSI + Cellular MESSAGE – Field Units Imperial • GUSTO / Electrochemical + wireless ranging + WiFi XML Data Transport p Format Common Data transmission format Investigating other forms of XML serialisation for increased efficiency <?xml version="1.0" encoding="UTF-8"?> <sd:DataPacket xmlns:sd="http://www.message-project.org/xml/2007/03/sensordata" > <sensorID>d8091be8-1513-4b89-9f9e-8391ba34bd10</sensorID> <packetOrder>-1</packetOrder> <timestamp>2007-05-22+01:00</timestamp> <location> <latitude>51.51526</latitude> <longitude>-0.14200687</longitude> <height>-1.0</height> g g <bearing>-1.0</bearing> <speed>-1.0</speed> <quality>0.0</quality> </location> <pollutionData> <CO>23.0</CO> <PM10>0</PM10> <PM2_5>0</PM2_5> <PM1>0</PM1> <PM0_1>0</PM0_1> PM0 1 0 /PM0 1 <NO>0</NO> <NO2>0</NO2> <NOx>0</NOx> <O3>0</O3> <SO2>0</SO2> <CO2>0</CO2> <noise>0</noise> </pollutionData> </sd:DataPacket> Data Formats - UTMC Schema Source: Visalakshmi Suresh, MESSAGE Project, North East Regional e-Science Centre e-Science Platform – Data Capture Th Three Layer L Architecture A hit t • Application Layer • Realtime Data Layer • Sensor Layer e-Science Platform – Analysis Data Analysis Real time data analysis requires the balancing of a number of factors to identify the available capacity for carrying out real time data analysis: y Algorithms y Local/Remote processing power y Data volumes y Communications bandwidth y Cost Visualisation Making live/historic data available for analysis by individuals (e.g. transport planners, scientists, i ti t etc.) t ) • Many visualisation platforms/APIs available • Real time integration with third party data e-Science Platform – Analysis In-network Analysis Where computational power and communications allow, in-network processing and analysis of data presents many opportunities. MESSAGE is prototyping a system to demonstrate this: • Peer-to-Peer (P2P) data sharing between local devices • Use of K-means clustering algorithm between a group of nodes • Potential for hotspot identification, local dispersion prediction, … • Event triggers sent to central platform platform. Local Processing Platform ARM-based processing nodes running Linux. 802.11 WiFi (ad-hoc), Bluetooth, GPS, USB connectivity. Varietyy of software and interfacing g options. p Field Trials – South Kensington DUVAS sensors Route A DUVAS sensors Route B DUVAS sensors – static Newcastle MOTEs – static Cambridge sensors – cyclists/pedestrians Q / / coAQM/DUVAS/MOTE located deployment Field Trials Early testing is underway Earl nder a for a co co-ordinated ordinated field trial that will be carried out in South Kensington to demonstrate the capture, streaming and processing of pollution data from different sensor types into a co-ordinated platform for analysis and visualisation. OGSA-DAI/DQP in MESSAGE End-user/application queries OGSA-DQP OGSA-DAI OGSA-DAI OGSA-DAI DB DB DB London Newcastle Cambridge Data from sensors Data from sensors Data from sensors THANK YOU! jeremy.cohen@imperial.ac.uk robin north@imperial ac uk robin.north@imperial.ac.uk www.message-project.org j