Collaborative Semantic Mapping for Updating the Digital Twin in controlled Indoor Environment
Article : Articles dans des revues internationales ou nationales avec comité de lecture
Efficient management of indoor spaces is increasingly critical for applications such as security, evacuation planning, and roboticdeployment. Digital twin technology has emerged as a transformative solution, providing a real-time link between the physicalenvironment and its virtual counterpart to enable monitoring, simulation, analysis, and performance optimization. This paperintroduces a novel collaborative approach to semantic mapping that enhances digital twins with semantic maps enriched withcontextual information about the environment. Designed specifically for controlled indoor settings, the approach assumes theavailability of prior knowledge through 3D CAD models and a managed environment. The proposed method leverages a state-of-the-art single-robot semantic mapping technique to collect semantic information using an RGB-D camera, integrating objectdetection, scene segmentation, and computational geometry to generate detailed point clouds and define object occupancyzones. Building on this foundation, a collaborative framework is developed for maintaining and updating the semantic mapwithin the digital twin. Autonomous mobile robots generate individual semantic maps, which are communicated to the digitaltwin and incrementally integrated into the existing map. The framework employs spatial and semantic correspondences,along with prior knowledge from the digital twin, to merge and synchronize asynchronous data collected by multiple robots.This process addresses challenges such as inaccurate object representations, class ambiguities, and data overlaps, while alsocapturing gradual or occasional environmental changes to ensure the digital twin accurately reflects real-world conditions.Experimental evaluations in representative office environments demonstrate the method’s effectiveness for scenarios involvingmoderate structural evolution, such as adding, moving, or deleting known objects. The results validate the framework’s practicalrelevance within its defined scope.