Sidewalk Condition Assessment Leveraging Machine Learning/AI and Mobile LiDAR

This research project seeks to demonstrate the feasibility of mobile LiDAR data as a cost-effective means to support efficient inventory and condition assessment of sidewalks at DDOT. The objective of this research project is twofold: 1) to develop and validate an improved point cloud data processing method that can automatically map pedestrian infrastructures (e.g., sidewalk, curb ramp, etc.) using point cloud data, which will demonstrate its feasibility for network-level analysis (e.g., using Ward 7 as the pilot testing field; 2) if it proves feasible, to apply the developed, automated method to the city-wide mobile LiDAR dataset collected by CycloMedia and generate a complete pedestrian infrastructure map in the entire District. If successful, the outcome of the proposed method in this research project will layout a solid framework toward a sidewalk assessment management program that routinely and cost-effectively inspects, evaluates, and maintains pedestrian infrastructure in the entire District.

    Language

    • English

    Project

    • Status: Programmed
    • Funding: $200000
    • Sponsor Organizations:

      District Department of Transportation

      250 M Street, SE
      Washington, DC  United States  20003
    • Managing Organizations:

      Howard University

      Transportation Research Center
      2366 Sixth Street, NW
      Washington, DC  United States  20059
    • Performing Organizations:

      University of Massachusetts, Amherst

      Baystate Roads Program, 214 Marston Hall
      Amherst, MA  United States  01003
    • Principal Investigators:

      Ai, Chengbo

    • Start Date: 20220701
    • Expected Completion Date: 20231231
    • Actual Completion Date: 0

    Subject/Index Terms

    Filing Info

    • Accession Number: 01845187
    • Record Type: Research project
    • Source Agency: District Department of Transportation
    • Files: RIP, STATEDOT
    • Created Date: May 12 2022 2:57PM