In this project, traffic flows in the 5- to 30-minute intervals are predicted for improving the conveyed messages in the highway VMSs which leads to more acceptable traffic management
- The traffic flow rate was measured in several highways of the city of Isfahan using data gathered from the installed cameras. Then, the noise was separated from the main signal through the wavelet transformation. Traffic flow was predicted for different time intervals by the neural network at error rate of less than 5%. Accuracy of the outputs was remarkably higher than the conventional models without noise separation. To validate the proposed method, the US highway standard data was further predicted within the same error interval.
Municipality of Isfahan
Isfahan University of Technology
Dr. Meisam Akbarzadeh