Automated Sidewalk Quality and Safety Assessment System
PI: Randall Guensler, Ph.D., Georgia Institute of Technology
The research team will calibrate and field-deploy a new automated sidewalk quality assessment tool (Adroid application) recently developed at the Georgia Institute of Technology. The application runs on a Toshiba Thrive tablet attached to the base of a standard, low-cost wheelchair. The user pushes the wheelchair along a sidewalk and the tablet collects GPS position data, vibration data (from accelerometers), and a high-resolution video on a SD memory card. When the SD card is activated on a personal computer, field data are transferred via the Internet to the Georgia Tech server for automated post-processing. By adapting mapping and video processing tools previously developed by for vehicle tracking and pothole identification, the system will process the GPS and video data to: 1) create GIS-based base sidewalk inventories, 2) automatically estimate sidewalk width, 3) record the localized presence of walkway obstructions, and 4) visually identify major sidewalk cracks requiring maintenance.
An expert system will use the accelerometer and video data to identify sidewalk discontinuities that do not meet Americans with Disabilities Act (ADA) design requirements and classify sidewalk surface repair needs. In consultation with local transportation planners and local public interest groups, the team will calibrate the expert system and develop a sidewalk quality index (SQI) to prioritize sidewalk repairs and improvements. For example, the system can assign the highest repair priority to sidewalks that do not conform to ADA requirements (improving safety and reducing the potential for law suites) and lower repair priority to sidewalks with simple surface roughness defects.
Georgia Tech's open source mapping applications on the research web server will facilitate public access to the spatial data and SQI results. The team will collect four months of sidewalk data along critical transportation and pedestrian corridors identified through stakeholder input and review of pedestrian incidents database. The researchers will coordinate data collection in Atlanta Georgia and will collaborate with another partner university to collect data in another city and state. The low-cost system will facilitate widespread data collection efforts led by volunteers from public interest groups (e.g. PEDS), educational institutions (e.g. universities/high-schools), and local stakeholders (e.g. neighborhood planning unit participants).