Purpose of the Pilot
The main purposes of the pilot are:
- Create a traffic flow prediction model
- Study the data in a research perspective
- Collect a history database about traffic flow data
- Investigate what kinds of algorithms are possible to use for this specific pilot
- Investigate the key factors that might be of value and have an impact on the Gothenburg pilot.
Idea and Motivation
The capacity of the roads in many cities has not grown as rapidly as the traffic flow has increased and this requires big efforts to meet the traffic needs now and in the near future.
To test the ideas, Gothenburg is proposed as a ROADIDEA pilot, due to data access from several sources in the area and also that both the Swedish ROADIDEA partners, Semcon and Klimator, are located there. In practical case the pilot uses data from Vägverket to make traffic flow predictions.
Gothenburg’s traffic situation is quite specific since the city is separated by a river and there are only three connections over the river close the city (figure 11), the last one built forty years ago. Since then, more roads have been built in Gothenburg but no more connections over the river. This causes a lot of traffic problems and might be typical for other cities with growing transport needs as well.
Fig. 11 Map of Gothenburg.
The key research question of the pilot is how the use of modern technology can result to a better traffic situation when making analysis with historical data. A very important part of the pilot is to evaluate the traffic flow prediction model.
Pilot Description
In this pilot a traffic flow prediction model, process and XML interface to UI will be created and demonstrated. The model will use both historical model data and real time data (like traffic density and traffic speed). Real time data will be used on a specific algorithm and as a result will be current traffic flow and a prediction of the future. The model uses traffic flow information that is collected to historical database together with real-time calendar data. This data is categorised into three different types of date: workdays, Saturdays ands holidays. The data in these categories are re-categorised into half an hours time periods. These results are saved into static database that is used to compare with real-time traffic flow data. A more detailed description of the model can be found in the chapter 6.
The end user case of the pilot will be a mobile, browser-based website and/or mobile application which illustrate both the current traffic flow and traffic flow prediction. The application will be realised by Logica. Destia will support the process by serving all the input data that comes from the pilot. All the mobile application records will be saved to Logica’s database to be used e.g. in the predictions.
Data Sources
To build up this Pilot, several different data sources are going to be used and combined with new and improved algorithms.
• Traffic Information from Sensors and Cameras
Gothenburg has a lot of sensors and cameras to measure the traffic and include several parameters such as speed, flow, number of vehicles, distance between vehicles etc. These parameters are sampled in a period of one minute and stored in a database of the Swedish road administration. For this specific pilot one minute data have been requested for the whole period of a selection of important gateways of the roads with high amount of traffic.
• Road Weather Information
VVIS (Road Weather Information System) is the Swedish Road Administration’s system to measure weather on the Swedish roads. There are about 700 stations around the country to measure road and air temperature, dew point, air humidity, precipitation, wind speed and wind direction etc.
Pilot Data Period
The pilot will use historical data from the period of 1st of November 2007 to 31st March 2008. For this period, Semcon and Klimator have lot of historical data to use as a base for the pilot.
Innovation value of the pilot
The main innovation of the pilot is that it will produce traffic flow predictions that can be used in traffic related applications which help road users to plan their every-day trips; e.g. time and route of travel.
The Challenge for Innovation
Prediction of the future is always challenging because of the unlimited number of factors that can affect the final result. The challenge of this pilot is real-time prediction performance because there are huge amounts of historical data.
Use case of the potential pilot service
The potential use cases for the pilot service are all portable devices that can be used in traffic. This includes personal navigation devices, mobile phone applications, in-vehicle computers etc.
Necessary ROADIDEA Partners
The idea is that the Gothenburg pilot should be done in cooperation with the following partners in WP6:
• Destia
• Logica
• Semcon
• Klimator
Estimated Pilot Schedule
Q2 2009: Traffic flow data collection from Vägverket starts.
Q3 2009: Traffic flow prediction processes are made.
Q3 2009: Prediction XML interfaces are made.
Q3 2009: Mobile application is ready for demo use.
Q4 2009: Demonstration of the pilot