An Artificial Intelligence (AI) Traffic Data Analysis Tool for Advanced Freeway Traffic Management

An Artificial Intelligence (AI) Traffic Data Analysis Tool for Advanced Freeway Traffic Management 3/20/2018

Planned or unplanned traffic events cause various magnitude of traffic congestion and safety impact to road users. It was estimated that 60% of congestions are non-recurrent and 15% of incidents are secondary to the primal accidents. Thus local or regional Traffic Management Centers (TMCs) have spent tremendous labor and budget resources detecting and responding to these events. In this project, we propose to develop, deploy, and test an Artificial Intelligence (AI) based Traffic Data Analysis Tool (AI-TDA) for Advanced Traffic Management decision support. This tool, which will be built upon our existing Traffic Signal Operation Analysis Expert System (SOES) software, leverages most recent advances in AI and ML such as Convolutional Neural network (CNN), Long Short-Term Memory (LSTM) network, video/imagery intelligence processing and exploitation, multi-source “big data” analytics, and decision recommendations/support for freeway and arterial traffic incident management applications. This tool, when fully developed, will be able to collect real-time field data and make recommendations to TMC operators/engineers in response to the traffic variations in the road network.