Simultaneous Localization and Mapping in A Camera Network

Zhiwei Liang,Xudong Ma,Xianzhong Dai

DOI: https://doi.org/10.1163/016918610x497107

IF: 2.057

2010-01-01

Advanced Robotics

Abstract:In this paper, we present a two-level mapping method that allows us to obtain accurate grid maps of large environments equipped with a camera network in real-time. The local map level is composed of a set of local maps, each of which is built depending on a robot's laser sensor and odometry using a Rao-Blackwellized particle filter method once the robot enters a new camera's visual field. Meanwhile, we also solve the camera calibration problem by using a marker attached to the robot that moves in a curve fashion in the camera's visual field. The global level is an adjacency graph whose arcs are labeled with the relative transformations between local maps. Among these relative transformations, of particular importance are the revisiting constraints. When the robot revisits a previous local map, it can be detected and relocated by a calibrated camera corresponding to the local map. Then a revisiting constraint can be attained and added into global constraints. To obtain an accurate and consistent global map, a stochastic gradient descent algorithm based on a simple state space is employed to optimize the existing graph after adding each global constraint. Experimental results carried out with a real robot in a large-scale indoor environment illustrate the advantages of our methods over two current state-of-the-art approaches. (C) Koninklijke Brill NV, Leiden and The Robotics Society of Japan, 2010