iRotate: Active Visual SLAM for Omnidirectional Robots

In this paper, we present an active visual SLAM approach for omnidirectional robots. The goal is to generate control commands that allow such a robot simultaneously localize itself and map unknown environment while maximizing the amount of information gained consuming as low energy possible. Leveraging robot’s independent translation rotation control, introduce multi-layered V-SLAM. top layer decides on informative locations generates highly paths them. second third layers actively re-plan execute path, exploiting continuously updated local features information. Moreover, two utility formulations account presence obstacles in field view location. Through rigorous simulations, real experiments, comparisons with state-of-the-art methods, demonstrate our achieves similar coverage results lesser overall entropy. This obtained keeping traversed distance up 39% shorter than other methods without increasing wheels’ total amount. Code implementation details are provided open-source, all generated data available online consultation.