No abstract

Utilization and generation of indoor maps are critical elements in accurate indoor tracking. Simultaneous Localization and Mapping (SLAM) is one of the main techniques for such map generation. In SLAM an agent generates a map of an unknown environment while estimating its location in it. Ubiquitous cameras lead to monocular visual SLAM, where a camera is the only sensing device for the SLAM process. In modern applications, multiple mobile agents may be involved in the generation of such maps, thus requiring a distributed computational framework. Each agent can generate its own local map, which can then be combined into a map covering a larger area. By doing so, they can cover a given environment faster than a single agent. Furthermore, they can interact with each other in the same environment, making this framework more practical, especially for collaborative applications such as augmented reality. One of the main challenges of distributed SLAM is identifying overlapping maps, especially when relative starting positions of agents are unknown. In this paper, we are proposing a system having multiple monocular agents, with unknown relative starting positions, which generates a semidense global map of the environment.

Visual Simultaneous Localization and Mapping (SLAM) has been used for markerless tracking in augmented reality applications. Distributed SLAM helps multiple agents to collaboratively explore and build a global map of the environment while estimating their locations in it. One of the main challenges in Distributed SLAM is to identify local map overlaps of these agents, especially when their initial relative positions are not known. We developed a collaborative AR framework with freely moving agents having no knowledge of their initial relative positions. Each agent in our framework uses a camera as the only input device for its SLAM process. Furthermore, the framework identifies map overlaps of agents using an appearance-based method.