Domestic service robots such as lawn mowing and vacuum cleaning robots arethe most numerous consumer robots in existence today. While early versionsemployed random exploration, recent systems fielded by most of the majormanufacturers have utilized range-based and visual sensors and user-placedbeacons to enable robots to map and localize. However, active range and visualsensing solutions have the disadvantages of being intrusive, expensive, or onlyproviding a 1D scan of the environment, while the requirement for beaconplacement imposes other practical limitations. In this paper we present apassive and potentially cheap vision-based solution to 2D localization at nightthat combines easily obtainable day-time maps with low resolutioncontrast-normalized image matching algorithms, image sequence-based matching intwo-dimensions, place match interpolation and recent advances in conventionallow light camera technology. In a range of experiments over a domestic lawn andin a lounge room, we demonstrate that the proposed approach enables 2Dlocalization at night, and analyse the effect on performance of varyingodometry noise levels, place match interpolation and sequence matching length.Finally we benchmark the new low light camera technology and show how it canenable robust place recognition even in an environment lit only by a moonlesssky, raising the tantalizing possibility of being able to apply allconventional vision algorithms, even in the darkest of nights.
展开▼
