Development of a Cloud-Based Traffic Diagnosis and Management Laboratory Based on High-Coverage ALPR

Abstract

As the traffic volume in urban areas continues to increase, traffic congestion results in longer travel time, lower reliability, and larger energy consumption. Traffic diagnosis and management strategies are generally considered the most cost-efficient approaches to reduce traffic congestion. Automated License Plate Reader (ALPR) is one of the most valuable data sources for traffic diagnosis and management. It outperforms many traditional methods of traffic data collection in terms of cost and accuracy. Many cities have deployed their ALRP systems in the past two decades. In those cities, traffic networks are highly covered by ALPRs. ALPRs produce a tremendous amount of data of moving vehicles at every moment. The information and insights hidden within the data can be a fortune for comprehensive traffic diagnosis and management. To fully take the advantage of ALPR data, more scalable and efficient data management, analysis algorithms, and visualization tools are required to process large-volume ALPRs data within reasonable time and budgets. An extensive body of research about ALPR data exists but mainly focuses on small-scale use of ALPRs on estimations of traffic flow characteristics. This dissertation plans to develop a traffic diagnosis and management laboratory utilizing high-coverage and large-scale ALPRs data. The online laboratory consists of modules for traffic condition reconstruction, diagnosis-oriented visual analysis, and traffic control and management decision support. This research contributes to (1) Designing a comprehensive and extensible framework for traffic diagnosis and management with ALPR data; (2) Developing highly scalable and efficient algorithms for traffic condition reconstruction; (3) Implementing and designing advanced visualization tools for traffic condition monitoring and diagnosis; and (4) Exploring the capability of utilizing ALPR data to improve the performance of data-driven traffic control and management strategies.