Traffic Surveillance Data Processing in Urban Freeway Corridors Using Kalman Filter Techniques

Details

• Creators:
  Houpt, Paul K. ; Athans, Michael ; Orlhac, Daniel G. ; Mitchell, William J.
• Corporate Creators:
  Massachusetts Institute of Technology
• Corporate Contributors:
  John A. Volpe National Transportation Systems Center (U.S.) ; United States. Dept. of Transportation. Research and Special Programs Administration
• Subject/TRT Terms:
  • Algorithms
  • Estimates
  • Freeways
  • Highway traffic
  • Highway traffic control
  • Human information processing
  • Methodology
  • Partial differential equations
  • Performance
  • Real time control
  More +
• Publication/ Report Number:
  DOT-TSC-RSPA-78-18
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States
• Corporate Publisher:
  United States. Dept. of Transportation. Research and Special Programs Administration
• NTL Classification:
  NTL-PLANNING AND POLICY-PLANNING AND POLICY ; NTL-PLANNING AND POLICY-Travel Demand ; NTL-PLANNING AND POLICY-Management Systems ; NTL-PLANNING AND POLICY-Access Management
• Abstract:
  Real-time surveillance of traffic conditions on urban freeway corridors using spatially discrete presence detectors is addressed. Using a finite-dimensional (macroscopic) fluid-analog model for freeway vehicular traffic flow, an extended Kalman filter is proposed as a data-processing algorithm to obtain minimum variance estimates of spatial mean speed and density. It is shown that certain model parameters associated with available roadway capacity can be estimated on-line with a variation of the extended Kalman filter, and furthermore, that the time signatures associated with these estimates provide quantitative information concerning the presence of anomalous (i.e., incident) traffic events. Performance of the surveillance algorithm is evaluated using a detailed, multi-lane microscopic vehicle simulation which retains a stochastic mix of driver-vehicle types and passing. Complexity problems associated with the computer implementation of the extended Kalman filter are addressed, and techniques for decentralized realization are proposed.
• Format:
  PDF
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/11304/dot_11304_DS1.pdf
• File Type:
  [PDF - 3.13 MB]
• Collection(s):
  Volpe Center
• Main Document Checksum:
  urn:sha256:50d83d45fe7b09c228e5fafb8a23a68a34c632a9265e06c0dea09ecc0e26e37a