Robot-Based Uniform-Coverage and High-Resolution LIDAR Mapping for Physically-Grounded Metaverse Applications

Abstract

Connecting the tangible world bidirectionally with its digital counterpart is fundamental for Physically-grounded Metaverse Applications (PMAs). For instance, in visually-guided human assistance, space-time analytics, eXtended Reality (XR) or automation use cases, the ability to capture, model, and exploit dynamic environments is essential. PMAs unify tangible and digital premises through intelligent capabilities built on detailed, gap-free and up-to-date 3D models of buildings delivering dependable functionality with utility value. To create and democratize such Metaverse benefits, it is fundamental to enable non-experts to easily capture, annotate and exploit full-fledged 3D models of buildings. Therefore, open IoT-XR platforms, algorithms, tools, and compelling user experiences are needed. Key factors to address such production-level 3D mapping technologies are time, cost, instrumentation- and operational-complexity for generating high-resolution maps with uniform-coverage. This work introduces an 3D surveying method via a semi-automatic mobile robot-mapping system. The robotic platform and novel algorithms improve the mapping process by ensuring sampling regularity, increasing spatial coverage, rejecting outlier points, and reducing total mapping time. This article focuses on real-world challenges and lessons learned from a novel closed-loop volumetric scanning-and-analysis method. Extensive experiments support claims in large maps.