This project continues research aimed at real-time detection and use of pedestrian traffic flow information to enhance adaptive traffic signal control in urban areas where pedestrian traffic is substantial and must be given appropriate attention and priority. We focus on designing a universal method to estimate the pedestrian density and use it as a controlling signal for the adaptive real-time scheduler. Our recent work will result in an extended intersection scheduling procedure that integrates sensed pedestrians and vehicles into aggregate multi-modal traffic flows and allocates green time on this integrated basis.

We focus on building a universal algorithm to estimate the pedestrian density at intersections.  It is undesirable to spend lots of time to train or adjust the estimator. Contrarily, we hope it can plug-and-play.  Hence, we adopt a group of detection grids and record a series of triggered grid patterns as features for a learning algorithm instead of using complicate image processing techniques to get the estimation.

Project lead: Hsu-Chieh Hu

Project time period: Jan 2015-Dec 2015