Control and forecasting in transportation networks
Principal Investigators :
- Prof. Carlos Canudas de Wit, NECS project-team, Inria Grenoble Rhône Alpes
- Prof. Roberto Horowitz, University of California Berkeley
COMFORT addresses open issues for Intelligent Transportation Systems (ITS). The goal of ITS is to use information technologies (sensing, signal processing, machine learning, communications, and control) to improve traffic flow, as well as enhance COMFORT and safety of drivers. These tools are doubtlessly of very high societal and economical value. It has been established over the past several decades, through field studies and many scholarly publications, that ITS tools can significantly improve traffic flows on congested freeways and streets. COMFORT is concerned with the following three main objectives: (i) Model validation and robust modeling for traffic estimation, control and forecasting; (ii) New methods for traffic forecasting; and (iii) New methods for distributed traffic control and estimation.
The issue of maximizing the green-light interval that allows uninterrupted traffic flow along the entire arterial is nowadays critical. Benefits of traffic lights coordination are undeniable in terms of reducing delay, unnecessary stops, energy and fuel consumption, pollution. COMFORT has addressed this problem through the use of two control actions: traffic light offsets and variable speed limits. Specifically, the state of- the-art of the optimization problem was enriched in order to account for traffic energy consumption and network travel time, thus avoiding impractical or undesirable solutions. A traffic microscopic simulator was used to assess the performance of the proposed technique in terms of energy consumption, travel time, idling time, and number of stops. The correlation of theoretical bandwidth with known traffic performance metrics was studied, and an analysis of the Pareto optimum was carried out to help the designer to choose a trade-off in the multi-objective optimization problem. Finally, an evaluation of the traffic performance at different levels of traffic demand, aimed at showing the best operation conditions of the proposed strategy, and at devising a demand-dependent optimization, was performed. The main result of the work can be summarized as follows: At under-saturated traffic conditions (i.e. all queues at the traffic lights are dissipated within the cycle time), the variable-speed limits control is able to reduce the energy consumption by 35%, while ensuring no loss in terms of travel time.
Publications and Awards:
- 4 Conference papers.
- 2 Journal papers & 1 Journal paper (under review).
C. Canudas de Wit, F. Morbidi, L. Leon Ojeda, A. Kibangou, I. Bellicot, P. Bellemain. Grenoble Traffic Lab: An experimental platform for advanced traffic monitoring and control. IEEE Control Systems Magazine (2015).
More about COMFORT:http://necs.inrialpes.fr/v2/pages/comfort/EA_homepage_COMFORT.html