SMARTCITY TAIPEI

Status：Complete

In response to the promotion of the smart city initiative in Taipei, the Department of Transportation has introduced dynamic signs on some arterial road networks with big traffic flow fluctuations since 2019. Through real-time monitoring of traffic and standby vehicles, traffic signs at intersections can be dynamically adjusted to improve the continuous flow of traffic and shorten the peak period. However, the calculation of the time of dynamic signal systems in the past was often limited by the quality and the number of deployment of vehicle detectors (VD), which in turn affected the accuracy of traffic flow forecasts.

This project supplements the vehicle detectors with real-time road information and adopts the cloud system with AI algorithms to provide forecasts for road traffic flows and speeds. A visual city route network model is constructed with a deep learning algorithm to conduct the forecast of traffic flows and speeds in the next 2 hours. It is updated every 15 minutes to improve the performance of dynamic signal control. In the future, the connection of API signals can be used to expand the model to various scenario applications, such as extending the green lights for emergency rescue vehicles, reorganizing the time system, improving forecast for bus arrivals, and logistics scheduling.