Development and implementation of a real-time big-data management architecture for effective adaptive traffic signal control

Wuping Xin, Elena S. Prassas

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Adaptive traffic signal control dynamically adjusts traffic signals based on prevailing traffic conditions. The acquisition and processing of real-time traffic data play a crucial role. The emerging trend of "big data", characterized as "three Vs'", i.e., Volume, Velocity and Variety, potentially enables novel signal control concepts and more effective adaptive signal control implementations. However, there has been a lack of relevant real-time big-data management architecture - an architecture that recognizes the disadvantages of existing general-purpose big-data technologies such as Hadoop/MapReduce or NoSQL, an architecture that is specifically targeted and optimized for adaptive signal control, capable of managing big-data that is very large (volume), very fast (velocity), and diverse (variety), and an architecture that allows real-time predictive analysis and performs regional adaptive traffic control that calls for parallel executions of relevant signal optimization algorithms for different sub-areas of complex traffic networks. This paper first examines the historical evolvement of adaptive traffic signal control, and points out the challenges and opportunities in nowadays data-rich environment. The relevance of the generalpurpose big-data technologies (MapReduce/Hadoop and NoSQL) is discussed from the signal control perspective. A new real-time big-data management architecture is proposed, considering massive realtime traffic data available nowadays and new types of data emerging in future. These data are generally collected at high frequency, in large amount, and supplied from different sources in realtime. The proposed architecture is specifically targeted for adaptive signal control applications. It features a hybrid design with both centralized and distributed elements, taking into account the efficient data archival and retrieval at physical disk sectors and memory levels, real-time traffic data fusion and synthetizing, in-memory caching and indexing, and a set of customized analytics supporting the novel concept of Signal Optimization Repository in adaptive traffic control. This architecture has been implemented as the core technology of the ACDSS system, which is a multiregime, variable-objective adaptive traffic control system developed by KLD. A case study is presented showing the real-life application of the proposed architecture in ACDSS operations with hundreds of signalized intersections of New York City arterials and grid networks.

Original languageEnglish (US)
Title of host publication21st World Congress on Intelligent Transport Systems, ITSWC 2014: Reinventing Transportation in Our Connected World
PublisherIntelligent Transport Systems (ITS)
StatePublished - 2014
Event21st World Congress on Intelligent Transport Systems: Reinventing Transportation in Our Connected World, ITSWC 2014 - Detroit, United States
Duration: Sep 7 2014Sep 11 2014

Other

Other21st World Congress on Intelligent Transport Systems: Reinventing Transportation in Our Connected World, ITSWC 2014
Country/TerritoryUnited States
CityDetroit
Period9/7/149/11/14

Keywords

  • ACDSS
  • Adaptive traffic signal control
  • Big-data
  • Grid network
  • New York City traffic signal control
  • Real-time traffic control
  • Traffic management

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Mechanical Engineering
  • Automotive Engineering
  • Transportation
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Development and implementation of a real-time big-data management architecture for effective adaptive traffic signal control'. Together they form a unique fingerprint.

Cite this