Sidewalks are at the heart of walkable urban centers, offering a mode of travel that connects nearly all other travel options and community activities. However, sidewalks are not safe and accessible to all. Many individuals encounter sidewalk obstacles, like steep inclines, damaged concrete or curb barriers, that limit their access to the benefits of urban life. Though we cannot make all sidewalks accessible overnight, foreknowledge of avoidable obstacles helps promote more pedestrianism and use of public transport. Presently, there are no devices for cities to collect this data in consistent, scalable ways, nor are there data standards to enable sharing across cities and organizations. The OpenSidewalks Hardware project: Sidewalk Mapping on the Edge is a first, important step in this direction.
The Sidewalk Mapping on the Edge hardware project promotes equity in pedestrian routes by making sidewalks first-class members of an open data transportation network. When sidewalk data and sidewalk connectivity can be collected in a standardized way, the data can provide the foundation for a variety of essential use cases and downstream activities. For example, we've used such data to create automated pedestrian trip planning customized to individual abilities in accessmap.io. These data can also provide base data for first-responder curbside access information, or rich analytic tools for data-driven urban planning. Access to safe pedestrian routes is a critical challenge if we are to steward healthy, inclusive, and resilient cities around the globe.
Our project enables cities to automate the standardized sidewalk data collection process. With computer vision deployed onto the Jetson Nano, the system can take in street-view images, segment what infrastructure is in the scene, and generate accurate, sidewalk mapping information every single second. The device can be mounted unto a stroller, wheelchair, or a scooter, the standardized sidewalk, in GeoJSON format, will be consistently collected without any manual intervention or surveying.
Here is our demo:
Comments