An Artificial Intelligence (AI) Based System for Advanced Freeway Data Collection and Analysis

Planned or unplanned traffic events, such as work zones, collision accidents, sporting games and stormy/snowy weather, arise along our roadway systems and affect normal traffic operations. These anomaly events cause various magnitude of traffic congestion and safety impact to road users. Thus, local or regional Traffic Management Centers (TMCs) have spent tremendous amount of resources responding to these events by monitoring traffic in a region 24/7, identifying anomalies, providing real-time traveler information, and executing and coordinating traffic control solutions. We proposes to further develop, deploy, and test an Artificial Intelligence (AI) based Traffic Data Analysis Tool (AI-TDA) for Advanced Traffic Management decision support. This tool, which builds upon our existing Traffic Signal Operation Analysis Expert System (SOES) software that was developed under the FHWA SBIR funding (Contract # DTRT5716C10004), leverages most recent advances in AI and ML such as Deep Neural network (DNN), Long Short-Term Memory (LSTM) network for traffic flow prediction, video/text data processing and exploitation, multi-source “big data” analytics, and decision recommendations/support for freeway and arterial traffic incident management applications. It will also apply to a signalized intersection for dilemma zone warning and signal operation optimization as a show case for AI/ML.