Map-based navigation for a mobile robot with omnidirectional image sensor COPIS

A Cooperative RSTA Research Program in Software and Intelligent Systems

“…The effect of noise on computing a robot's position by measuring the angles between known landmarks is investigated in [2]. Navigation and map building with a mobile robot using a conical mirror is considered by Yagi et al in [19] and [18].…”

“…Much of this work has been focused on designing sensors with a panoramic or wide field of view (see [16], [10], [13], [17], [9], [7] [11], [18] [3], [19], [15], [12], [14], [4], [8], [1], [6]). …”

Complex Analysis for Reconstruction from Controlled Motion

“…Recently, a number of approaches to navigation and reconstruction using omnidirectional systems have been proposed (Nayar [6], Svoboda [11], Onoe [7], Srinivasan [2], Yagi [12], Medioni [10]). The work by Yagi and Medioni is very similar but uses an already known environmental map.…”

Simultaneous Localization and Mapping with Stereo Vision

“…Observed lines are used to reduce odometric uncertainty using an extended Kalman filter (EKF), then the observations are in turn used to update an environment map containing 2D point features representing the observed vertical lines. Yagi, Nishizawa, and Yachida's system [21] took a similar approach but used a single omnidirectiona vision sensor and accumulation of measurements over time, rather than stereo, to determine the positions of vertical line landmarks. These systems and others have amply demonstrated the efficacy of VSLAM based on line landmarks in constrained indoor environments with smooth floors.…”