A new approach to occupancy map modeling in navigation assisted by computer vision

Abstract

[EN] Occupancy grid maps have been widely used in robotics for navigation assistance. Most of the current occupancy maps use grids based on Cartesian coordinates. In this paper, we introduce a new approach to occupancy map modeling which represents the grid using cylindrical coordinates instead of Cartesian coordinates. The new approach is applied to create a 3D model of the surrounding space. Several computer vision systems and sensors including a stereo vision system, a laser-based depth system and some movement and position sensors are considered to be used to provide the inputs. Simulations are provided to show that the proposed approach presents some improvements over Cartesian coordinates based model. © 2011 CRL Publishing Ltd.