In field or indoor environments it is usually not possible to provide service robots with detailed a priori environment and task models. In such environments, robots will need to create a dimensionally accurate geometric model by moving around and scanning the surroundings with their sensors, while minimizing the complexity of the required sensing hardware. In this work, an iterative algorithm is proposed to plan the visual exploration strategy of service robots, enabling them to efficiently build a graph model of their environment without the need of costly sensors. In this algorithm, the information content present in sub-regions of a 2-D panoramic image of the environment is determined from the robot's current location using a single camera fixed on the mobile robot. Using a metric based on Shannon's information theory, the algorithm determines, from the 2-D image, potential locations of nodes from which to further image the environment. Using a feature tracking process, the algorithm helps navigate the robot to each new node, where the imaging process is repeated. A Mellin transform and tracking process is used to guide the robot back to a previous node. This imaging, evaluation, branching and retracing its steps continues until the robot has mapped the environment to a pre-specified level of detail. The effectiveness of this algorithm is verified experimentally through the exploration of an indoor environment by a single mobile robot agent using a limited sensor suite.
Service robots; visual mapping; selflocalization; information theory; Mellin transform
