Recently, efforts have been made to standardize signal phase and timing (SPaT) messages. These messages contain the current signal phase with a prediction for the corresponding residual time for all signalized intersection approaches. This information can thus be used for efficient motion planning, resulting in more homogeneous traffic flows and more uniform speed profiles. In this work, we propose a time series prediction framework using aggregated traffic signal and loop detector data. We develop a hybrid deep-learning model based on classification and regression tasks to predict residual times and compare results to standard machine learning baselines
Download the slides for this talk.Download ( PDF, 3025.56 MB)