文章基本信息

标题：Simultaneous localization and mapping for tracked wheel robots combining monocular and stereo vision
本地全文：下载
作者：Filipe Jesus ; Rodrigo Ventura
期刊名称：Journal of Automation, Mobile Robotics & Intelligent Systems (JAMRIS)
印刷版ISSN：1897-8649
电子版ISSN：2080-2145
出版年度：2013
卷号：7
期号：1
页码：21
出版社：Industrial Research Inst. for Automation and Measurements, Warsaw
摘要：This paper addresses an online 6D SL AM method for a tracked wheel robot in an unknown and unstructured environment. While the robot pose is represented by its position and orientation over a 3D space, the environ- ment is mapped with natural landmarks in the same space, autonomously collected using visual data from feature de- tectors. The observation model employs opportunistically features detected from either monocular and stereo vi- sion. These features are represented using an inverse depth parametrization. The motion model uses odometry read- ings from motor encoders and orientation changes mea- sured with an IMU. A dimensional-bounded EKF (DBEKF) is introduced here, that keeps the dimension of the state bounded. A new landmark classifier using a Temporal Dif- ference Learning methodology is used to identify undesired landmarks from the state. By forcing an upper bound to the number of landmarks in the EKF state, the computational complexity is reduced to up to a constant while not com- promising its integrity. All experimental work was done using real data from RAPOSA-NG, a tracked wheel robot developed for Search and Rescue missions.
关键词：simultaneous localisation and mapping; ex- ; tended Kalman filter; feature detector; inverse depth ; parametrization; landmark evaluation; temporal di.er- ; ence learning