1 Institut français des sciences et technologies des transports, de l aménagement et des réseaux Session 3 Big Data and IT in Transport: Applications, Implications, Limitations Jacques Ehrlich/IFSTTAR h/ifsttar Chair of PIARC Technical Committee 2.1 «Road Network Operation»
2 What is Big Data The term refers to the collection, exploration and analysis of large amounts of data. By "great", everyone hears what he wants. Depending of application it could be several million gigabytes (GB) to for a few thousand kilo bytes Big Data is becoming a major issue in ITS due to the fact that vehicles (cars, trucks, bus, trains etc) are themselves data sources and can/will soon communicate
3 Big Data and Connected Vehicle Modern vehicles produce large quantities of very accurate information Location, speed, driver action (on headlights, turn signals, windshield wipers, accelerator pedal, brake, ESP etc) These data can be easily collected on the internal network of the vehicle (CAN BUS) Up to now they are neither recorded nor transmitted (excepted for private fleet like taxi) Vehicle will be connected within Pre-deployment is ongoing : DRIVE C2X project soon complete, SCOOP project under construction Then vehicles are becoming «mobile sensors» «Probe vehicule», Floating Car Data
4 Probe vehicles for which application? Mobility Transport operator can better adapt offer to demand Road Network Operation Road operators can get an accurate view of the network status (accident, trafic jam, travel time) Road safety Driver can anticipate road difficulties and avoid accident. Depending on the road event they can adapt their behavior and modify their itinerary Insurance Insurers can offer the best rates to their customers depending on the way they use their vehicle : pay as you drive Car manufacturers Will get information about their clients : behaviors, needs, Will get information about their vehicles : reliability, failures prevention etc.
5 Application for a near future : road diagnosis Road diagnosis will be possible without addition of new sensors on the infrastructure Slippery road (rain, ice, skid resistance) Dangerous curve : speed not in adapted to road geometry Road surface degradation: asphalt degradation, rut, pothole Etc
6 Data qualiy Main issues (1/2) Accuracy, false data, missing data Data representativeness ti Amount of data required to describe an vent with a good probability of occurrence? Data mining How to extract information from Big Data set and transform it into an understandable structure for further use? Security How to be resiliant to attack, malice, spoofing?
7 Main issues (2/2) Privacy Data ownership Who is the data owner? The driver Acceptance by users Avoid mixing black box (to be used in case of accident) and dinformation provided dby floating car data Standardization Interoperability for data exchange and format. Stakeholder and value chain Who does what? Who pays for what? Who get money for what?
8 Ongoing discussion in Europe and worldwide ERTICO Probe Data Work Group Trilateral task force EU, US, JPN Autonomous vehicle Probe Data Etc.
9 Than you for your attention Jacques Ehrlich Ifsttar Bld. Newton Cité Descartes Champs sur Marne Marne-la-Vallée Cedex 2 France Tél. +33 (0)