A solution to the simultaneous robot localization and camera-network calibration problem of ubiquitous robot system

作者：Wu Peiliang^*; Kong Lingfu; Kong Liang

来源：ICIC Express Letters, 2011, 5(8 B): 2759-2765.

• Calibration problems
• Camera projection
• Dynamic environments
• Intelligent spaces
• Network calibration
• Networked robot
• Novel solutions
• Projection matrix
• Proposal distribution
• Rao-blackwellized particle fil-tering
• Robot interactions
• Robot localization
• Robot path
• Robot pose
• Robot system
• Simultaneous robot localization and camera-network calibration
• Ubiquitous computing technology
• Ubiquitous robot system

Due to the emergence of ubiquitous computing technology of Intelligent Space, a new class of networked robot called ubiquitous robot system has been introduced in the framework of Intelligent Space and robot interaction. Both the Intelligent Space sensor- network calibration and robot localization are two basic tasks in ubiquitous robot system. This article takes camera as sensor node of Intelligent Space sensor-network, and provides a novel solution to simultaneous robot localization and camera-network calibration problem for ubiquitous robot system. The joint posterior of robot pose and camera projection matrix are decomposed according to Rao-Blackwellized particle filtering theory, and the proposal distribution is established, from which the robot pose particles are sampled. The weight of each robot pose particle is obtained from both the controls of robot and observations of Intelligent Space camera-network. Simultaneously, the projection matrix of each camera node in the camera-network is calibrated according to both robot path and its corresponding imaging. The proposed method could obviously improve the precision of robot localization, especially in dynamic environment.

• 出版日期2011