Navigation Guided by Visual Markers and Path Planning for Mobile Robots in Confined Environments

In indoor environments without the possibility of geolocation, the safe navigation of an autonomous mobile robot is highly dependent on the sensors equipped in the mobile and the quality of the pose estimation performed. In turn, the path planning algorithm requires a suitable representation of the environment and an adequate estimation of the robot’s position. To this end, an autonomous navigation scheme with environment reconstruction is proposed. A data fusion algorithm allows obtaining an estimated pose of the vehicle with bounded error from the position recognition of fiduciary markers and a scanner-based SLAM algorithm. Included simulation results demonstrate the efficiency of visual marker recognition, position data fusion, and generation of safe path planning, considering that the environment has a non-regular surface.