You are here
Our research > Past Research Projects > Parametric and Nonparametric Multivariate Regression Models for Traffic Prediction and Control
Parametric and Nonparametric Multivariate Regression Models for Traffic Prediction and Control
- Project Team:
- Project Leader: Biswajit Basu and Prof. Margaret O’Mahony
- Description:
-
Prediction of traffic and other flow parameters on highways and freeways are important in terms of minimization of traffic delays lesser emission of pollutants and reduction of stresses to the drivers. Several parameters of interest such as flow, speed and capacity have been identified. The research focuses on the interrelationship of several parameters and aims at the development of comprehensive model to study their interrelated effects. Both parametric and nonparametric multivariate techniques would be used. In the parametric models, several statistical parametric estimation techniques would be used. These parameters will be used in the simulation of traffic data and to see how closely it fits to the real time traffic data. Traffic data from Dublin City transport would be used for this purpose.
Harmonic analysis of the data would be used for estimating the required parameters in frequency domain. Further, time frequency analysis techniques would be used for detecting the non-stationary of the data and build up of non-stationary models. Also, nonparametric techniques would be used for modelling. It would be compared to see if the model free estimation techniques would provide a better description of the traffic movement. Finally, the simulated traffic flow would be used to find the optimal control strategy in terms of movement of traffic flow. This would help in the design of traffic signal systems and to design the strategy in terms of optimality of the pedestrian and vehicular movement.
- Funding Agency:
