Accurate sidewalk map data are scarce. Sidewalk high-definition maps have been virtually non-existent. In a recently published article ("Closing the gap in sidewalk data") by our co-founder and CEO Dr. Meead Saberi discusses some of the gaps in sidewalk data for city planning as well as automated sidewalk robot delivery applications.
At footpath.ai, we view sidewalks as complex multi-layer environments with a graph layer as the base, a visual layer, and a semantic layer on top. Sidewalks are fascinating environments often governed by self-regulated actions of pedestrians, cyclists, scooters, and even pets. Unlike roads with pre-determined traffic lanes and directions, pedestrians move in multi-directions in a self-organized way.
Sidewalk delivery robots to safely and fully autonomously navigate from A to B require prior-built maps of sidewalks that can help them with improved localization, scene understanding, route planning, and motion planning.
Read more about gaps in sidewalk map data here: https://medium.com/@meead.saberi/closing-the-gap-in-sidewalk-data-a265a0e5432