Building maps for mobile robots by observing walking people

This paper presents a vision-based method for mobile robots to detect collision-free paths by observing walking people. The method (1) extracts person´s regions from each image frame by background subtraction, and (2) recovers paths from the trajectories of persons in successive frames. Due to the uncertainties inherent in the processes, there are many possible interpretations of the trajectories. To cope with the problem, recovered trajectories are classified according to their reliabilities, and recorded on a set of maps with different levels of reliabilities, Hierarchical Reliability Maps (HRM). We will show low reliability maps of HRM describe many paths roughly, while high reliability maps describe only reliable paths in detail.