Characterizing Pedestrian Network from Segmented 3D Point Clouds for Accessibility Assessment: A Virtual Robotic Approach

This study introduces a novel virtual robotic approach for automated characterization of pedestrian network accessibility from semantically segmented 3D LiDAR point clouds. With approximately 8 million Canadians living with disabilities, scalable accessibility assessment methods are critical. The proposed methodology integrates a Tangent Bug navigation algorithm—extended from 2D to 3D point cloud environments—with a triangular virtual robot grounded in ADA and IBC accessibility standards. The robot navigates classified point cloud data to simultaneously extract related parameters per step including those related to the accessibility assessment, including running slope, cross-slope, path width, surface type, and step height, aligned with the Measure of Environmental Accessibility (MEA) framework. Unlike existing approaches, the method characterizes not only formal sidewalk segments but also the critical transitional linkages between building entrances and the pedestrian network. Rather than evaluating features against fixed binary thresholds, it records continuous raw measurements enabling personalized accessibility assessment tailored to individual user profiles. Quantitative validation demonstrates high accuracy for path width (NRMSE = 2.71%) and reliable slope tracking. The proposed approach is faster, more cost-effective, and more comprehensive than traditional manual methods, and its segment-independent architecture makes it well-suited for future city-scale deployment.