AutoAlert: Automated Acoustic Detection of Incidents

Details

• Alternative Title:
  IDEA PROJECT FINAL REPORT, AUTOALERT: AUTOMATED ACOUSTIC DETECTION OF INCIDENTS
• Creators:
  Whitney, David A. ; Pisano, Joseph J.
• Corporate Contributors:
  TASC, Incorporated
• Subject/TRT Terms:
  Automatic Incident Detection Incident Detection Intelligent Transportation Systems
• Resource Type:
  Research Paper
• TRIS Online Accession Number:
  00734675
• Corporate Publisher:
  United States. Federal Highway Administration
• NTL Classification:
  NTL-INTELLIGENT TRANSPORTATION SYSTEMS-INTELLIGENT TRANSPORTATION SYSTEMS
• Abstract:
  This Innovations Deserving Exploratory Analysis (IDEA) project included the design and preliminary evaluation, as well as feasibility demonstration, of AutoAlert, an acoustic traffic sensor system that applies new signal processing algorithms to passive acoustic data to advance the state of practical acoustic incident detection techniques. These techniques, originally developed for national defense applications, will perform reliable, automatic, nearly instantaneous, all-weather incident detection under highly variable traffic conditions. Effective operation of urban high-capacity Intelligent Transportation Systems (ITS) requires speedy detection of incidents at chokepoints, such as tunnels, bridges and other aerial structures, and dense urban arterials. AutoAlert overcomes shortcomings of loop and video detectors, such as their inability to distinguish between incidents and congestion, and the need for a human-in-the-loop for video detection. The AutoAlert processor "hears" an incident before congestion builds, and can be used either as an independent detector, or its outputs can be combined (data fusion) with other detector outputs for joint improved decisions and incident verification. AutoAlert algorithms will provide a new level of incident detection timeliness and reliability (low false alarms) by applying sophisticated statistical models: Hidden Markov Models and Canonical Variates Analysis. These are used to analyze both short-term and time-varying signals that characterize incidents. 58 p.
• Format:
  PDF
• Collection(s):
  US Transportation Collection
• Main Document Checksum:
  urn:sha256:dc4fa7c40bb127d4620eee8d2063a60bb1d6ddb47770076f20c3d3ba7c9fad63
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/39341/dot_39341_DS1.pdf
• File Type:
  [PDF - 1014.12 KB ]