Simultaneous Localization and Mapping (SLAM) is crucial technology to enable vehicles to move autonomously. In real-world situations, large-scale environments are prone to encountering with the challenges of high time cost generally. Furthermore, when dealing with largescale mapping tasks, existing SLAM algorithms suffer from the difficulties such as high computational complexity, significant memory consumption and substantial cumulative drift. To enhance the mapping efficiency and accuracy of vehicles, this paper proposes a topological feature based multi-vehicle cooperative mapping method tailored for large-scale environments. By the method, the master vehicle conducts preliminary fusion through the first-round feature matching. By clustering the boundary points in the preliminarily fused map to obtain the exploration points for the vehicles, an explorationpoint model containing topological features is established to construct a cost function for the exploration points. Then appropriate exploration points are assigned to corresponding vehicle to perform next round exploring task, and each vehicle updates its local map according to the assigned exploration tasks and shares it with the master vehicle. Finally, the master vehicle calculates the attitude transformation between the local maps based on the exploration - point model, and emerge the localmaps to obtain the global map. Experimental results demonstrate that the method proposed can effectively improve both the efficiency and accuracy of mapping in large-scale environments.
Topological Feature Based Multi-Vehicle Cooperative Mapping System for Large-Scale Environments
16.05.2025
2793562 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Large-scale feature identification for indoor topological mapping
Tema Archiv | 2001
|LiDAR-Based Topological Mapping of Orchard Environments
Springer Verlag | 2022
|Compact and Efficient Topological Mapping for Large-Scale Environment with Pruned Voronoi Diagram
DOAJ | 2022
|