<title>Real-time map building for fast mobile robot obstacle avoidance</title>

Abstract

This paper introduces HIMM (histogramic in-motion mapping) a new method for real-time map building with a mobile robot in motion. HIMM represents data in a two-dimensional array (called a histogram gjid) that is updated through rapid continuous sampling of the onboard range sensors during motion. Rapid in-motion sampling results in a statistical map representation that is well-suited to modeling inaccurate and noisy range-sensor data. HIMM is integral part of an obstacle avoidance algorithm and allows the robot to immediately use the mapped information in real-time obstacleavoidance. The benefits of this integrated approach are twofold: (1) quick accurate mapping and (2) safe navigation of the robot toward a given target. HIMM has been implemented and tested on a mobile robot. Its dual functionality was demonstrated through numerous tests in which maps of unknown obstacle courses were created while the robot simultaneously performed real-time obstacle avoidance maneuvers at speeds of up to 0. 78m/sec.