Sampling-based floor plan analysis on BIMs

Abstract

This contribution explores a novel combination of methods from robotics and BIM by working on Building Information Models (BIMs) with the Open Motion Planning Library (OMPL), an extensible programming interface and reference collection of various algorithms for sampling-based motion planning. We present a seamless workflow from industrial BIM generation to sampling-based planning of collision-free paths for mobile robots on 3D floor plans extracted from BIMs. Beyond robot motion planning, we apply the workflow to generate huge amounts of paths for floor plan analysis. We assume the paths to resemble walking patterns where many paths crossing a given floor tile match with heavy usage of this space. The resulting paths are sampled in heatmaps, where pixels then correspond to floor tiles and the color value corresponds to the number of crossing paths, yielding a semi-automatic, easy-to-use method to analyze floor plan layouts in different scenarios.