A Crowdsourced Mapping of Sidewalk Accessibility Using Smartphones and Multimodal AI
Sidewalk condition is a key factor for determining whether a city is truly accessible.
Yet most cities lack reliable data. Existing methods of sidewalk monitoring are costly and slow, while satellite or vehicle-captured imagery misses the critical pedestrian perspective.
We propose a new method: Open Walks.
A web app that runs directly in any smartphone browser, bringing sidewalk-accessibility mapping to the hands of the community.
Albany St. Cambridge, MA
Each segment represents 3 meters of sidewalk.
Each step counts in sidewalk analysis. With Open Walks, each step becomes a data point containing:
Building on our existing Sidewalk AI model, we automatically analyze geolocated surface materials, obstacles, and cracks.
A Vision Language Model (VLM) then examines each sidewalk data point to extract additional qualitative accessibility information.
Together, all of these dimensions are combined into an accessbility score tailored to local standards.
Creating a powerful and scalable platform that reveals the full picture of sidewalk accessibility.
With more than 7.3 billion active smartphones worldwide, every scan contributes to a global, collaborative map of sidewalk accessibility.
Contribute to the mission and start mapping your city.
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Duis ut imperdiet felis. Aliquam erat volutpat. Nunc ullamcorper nibh a consectetur pharetra. Vivamus sit amet feugiat nulla. Duis pulvinar lorem sed suscipit sagittis. Class aptent taciti sociosqu ad litora torquent per conubia nostra, per inceptos himenaeos. Proin elementum tortor velit, quis dapibus dolor pretium at. Donec quis est sit amet metus dapibus rhoncus a vehicula arcu. Aliquam erat volutpat.
Morra, D., Mazzarello, M., Ratti, C. (2026). Open Walks: ..... . Journal Name.
The material on this website can be used freely in any publication provided that:
For more information,
senseable-contacts@mit.edu