A Weather-Responsive Automatic Incident Detection Algorithm with Generalized Detection Thresholds

A weather-responsive automatic incident detection (AID) algorithm with generalized detection thresholds is proposed for incident detection under both no rain and rain conditions. The proposed algorithm adopts adaptive detection thresholds generated by a generalized detection threshold function in which both pre-incident traffic flow and rain conditions are considered. As incident-induced changes in traffic conditions vary under different pre-incident traffic flow and rain conditions, consideration of pre-incident traffic flow and the condition of rain can contribute to improvements in the overall performance of AID algorithms. The discrete detection thresholds adopted by previous flow-based AID algorithms have a boundary value jump problem. In the proposed algorithm, a detection threshold function is adopted to generalize the relationship between the threshold and the pre-incident traffic flow. The detection threshold under any specific pre-incident traffic flow values can be obtained by the generalized detection threshold function. Traffic data, accident data and rainfall data were collected on a selected road network in Hong Kong to calibrate and validate the proposed algorithm. The proposed algorithm is then deformed into several algorithms, in which either one, or both of the pre-incident traffic flow and rain conditions are excluded in detection threshold determination. The proposed algorithm was found to perform better than these deformed algorithms. It is expected that the proposed algorithm could be applied to roads in cities with similar substantial rainfalls as experienced in Hong Kong.

• Supplemental Notes:
  • This paper was sponsored by TRB committee AHB20 Freeway Operations.
• Corporate Authors:

  Transportation Research Board

  500 Fifth Street, NW
  Washington, DC  United States  20001
• Authors:
  • Li, Xiangmin
  • Lam, William H K
  • Tam, Mei Lam
• Conference:
  • Transportation Research Board 93rd Annual Meeting
  • Location: Washington DC
  • Date: 2014-1-12 to 2014-1-16
• Date: 2014

Language

• English

Media Info

• Media Type: Digital/other
• Features: Figures; References; Tables;
• Pagination: 16p
• Monograph Title: TRB 93rd Annual Meeting Compendium of Papers

Subject/Index Terms

• TRT Terms: Automatic incident detection; Crash data; Rainfall; Traffic data; Traffic flow
• Geographic Terms: Hong Kong (China)
• Subject Areas: Data and Information Technology; Highways; Operations and Traffic Management; I71: Traffic Theory; I72: Traffic and Transport Planning;

Filing Info

• Accession Number: 01518197
• Record Type: Publication
• Report/Paper Numbers: 14-1067
• Files: TRIS, TRB, ATRI
• Created Date: Mar 12 2014 9:33AM