The objective of this project is to develop a comprehensive framework with a set of models to improve multi-modal traffic signal control, by incorporating advanced floating sensor data (e.g. GPS data, etc.) and traditional fixed sensor data (e.g. loop detectors, etc.). Especially, we are interested in addressing the challenges of multi-modal signal control under non-recurrent conditions, such as traffic incidents and planned special events, since non-recurrent congestions usually account for more than 50% of the total congestions. In order to accomplish this goal, an 18-month project is defined in this proposal with a multidisciplinary team assembled with two PIs from transportation engineering and computer science, respectively. First, we will conduct a comprehensive interview with transportation professionals, who can bring up existing state-of-practice, open issues and future challenges in multi-modal traffic signal control. The results of this interview will be compared with the recent complete interview with police officers, who have real-world traffic control experiences under non-recurrent events (N. Ding, He, and Wu 2014). Second, a two-component traffic data analysis will be performed on a variety of multi-modal data sources, including passenger cars, transit buses, light rail and emergency vehicles as well as commercial trucks, bicycles, and pedestrians. One component is to fuse multi-source and multi-modal data sets and predict the traffic state in near future. The other one is to identify the anomaly condition in traffic network, caused by traffic incidents (e.g., collisions, disabled cars, hazard materials, etc.) or special events (e.g., football game, parade, marathon, etc.). Third, multi-modal signal control algorithms will be developed to leverage the results derived from traffic data analysis, under both recurrent and non-recurrent congestion conditions. Finally, the proposed framework will be evaluated by microscopic simulation VISSIM and externally developed signal control modules. With consideration of advanced multi-modal and multi-source data, this research closely aligns with UTRC’s Focus Area #4: System modernization through implementation of advanced and information technologies as described in the RFP. Through alleviating traffic congestion and improving safety of the highway system, this work will also contribute to UTRC consortium’s themes in Economic Competitiveness and Livable Communities. The project team will work closely with Niagara International Transportation Technology Coalition (NITTEC), the City of Buffalo, and New York City on how the proposed algorithms and models could help in the development of a multi-modal traffic management and operations Decision Support System (DSS). The results will be disseminated to transportation authorities through webinars or workshops for workforce training.

Related Publications:

Cai, Y., H., Tong, W., Fan, P., Ji, and Q. He, “Facets: Fast Comprehensive Mining of Co-evolving High-order Time Series”, 21th ACM SIGKDD Conference on Knowledge Discovery and Data Ming (KDD) August, 2015. (acceptance rate: 159/819 = 19.4%)

Zhang, Z.*, Q. He, H. Tong, J. Gou, and X. Li, “Spatial-Temporal Traffic Pattern Identification and Anomaly Detection in a Large-Scale Urban Network”, Submitted to Transportation Research Part C

Su, X.*, Caceres, H.*, Tong, H., and Q. He, “Online Travel Mode Identification using Smartphones with Battery Saving Considerations”, accepted to IEEE Transactions on Intelligent Transportation Systems