pacific nw transportation consortium region 10 Pactrans

Combining Crowdsourcing and Machine Learning to Collect Sidewalk Accessibility Data at Scale

PI: Jon Froehlich (UW), jonf@cs.washington.edu, ORCID: 0000-0001-8291-3353

Co PIs: none

AMOUNT & MATCH: $50,000 from PacTrans; $50,000 Match

PERFORMANCE PERIOD: 8/16/2019 – 8/15/2021

STATUS: Completed

CATEGORIES: Sidewalks, Pedestrians, Accessibility, Crowdsourcing, Machine Learning

RESEARCH PROJECT HOT SHEET:

UTC PROJECT DOCUMENTATION:

FINAL PROJECT REPORT:

PROJECT DATA:

DESCRIPTION: Sidewalks significantly impact the mobility and quality of life of millions of Americans. In the proposal, we described new, scalable methods for collecting data on sidewalk accessibility using machine learning, crowdsourcing, and online map imagery as well as new interactive visualizations aimed at providing novel insights into urban accessibility.

As with our prior research, we will work closely with key stakeholders, including local governments and transit departments, mobility-impaired individuals and caretakers, and walkability advocates to help shape and evaluate the design of our tools.

While our proposed techniques and tools should work anywhere with OpenStreetMaps and available streetscape imagery (e.g., Google Street View, Mapillary), two of our three immediate deployment targets are cities in the PacTrans region: Newberg, OR and Seattle, WA.