Advanced Learning Algorithms for the Proactive Infrasonic Pipeline Evaluation Network (PIGPEN) Pipeline Encroachment Warning System

This research will advance development of self-training algorithms supporting seismic sensor systems that provide real-time warning of unauthorized right-of-way encroachment and excavation activity near a pipeline. he outcome will enable the sensor system to optimize its intruder detection algorithms based on learned characteristics of its local environment. Field tests are expected to demonstrate better than 97% alarm reliability with few false alarms. Physical Sciences Inc. (PSI), with American Innovations Ltd. (AI) and NYSEARCH, are addressing the technology gap of Early Warning Damage Prevention Monitoring Systems, specifically Advanced Development of Algorithms for Detecting Digging Threats and Avoiding False Alarms. This research will implement and evaluate self-training algorithms in the Proactive Infrasonic Gas Pipeline Evaluation Network (PIGPEN) autonomous distributed seismic sensor system. PIGPEN provides real-time warning of unauthorized right-of-way encroachment and excavation activity near a pipeline. Early warning enables a response to the potential intrusion in time to prevent pipeline damage, and thus preclude the additional cost and risk of repairs. The ideal PIGPEN alarm system would activate an intruder notification with 100% reliability and no false alarms. The project will enhance reliability by enabling PIGPEN to learn the characteristics of its local environment and optimize its intruder detection algorithms based on learned experience. Field tests are expected to demonstrate better than 97% alarm reliability with few false alarms.

Language

  • English

Project

  • Status: Active
  • Funding: $267000.00
  • Contract Numbers:

    DTPH56-10-T-000021

  • Sponsor Organizations:

    Pipeline and Hazardous Materials Safety Administration

    U.S. Department of Transportation
    East Building, 2nd Floor 1200 New Jersey Avenue, SE
    Washington, DC  United States  20590
  • Project Managers:

    Merritt, James

  • Performing Organizations:

    Physical Sciences Incorporated

    20 New England Business Center
    Andover, MA  United States  0810-1077
  • Principal Investigators:

    Fish, Michael

  • Start Date: 20100930
  • Expected Completion Date: 0
  • Actual Completion Date: 20120331
  • Source Data: RiP Project 31145

Subject/Index Terms

Filing Info

  • Accession Number: 01492394
  • Record Type: Research project
  • Source Agency: Department of Transportation
  • Contract Numbers: DTPH56-10-T-000021
  • Files: RiP
  • Created Date: Sep 7 2013 1:00AM