Uncertainty Analysis of a Landmark Initialization Method for Simultaneous Localization and Mapping

To operate successfully in any environment, mobile robots must be able to localize themselves accurately. In this paper, we describe a method to perform Simultaneous Localization and Mapping (SLAM) requiring only landmark bearing measurements taken along a linear trajectory. We solve the landmark initialization problem with only the assumption that the vision sensor of the robot can identify the landmarks and estimate their bearings. Contrary to existing approaches to landmark based navigation, we do not require any other sensors (like range sensors or wheel encoders) or the prior knowledge of relative distances between the landmarks. We provide an analysis of the uncertainty of the observations of the robot. In particular, we show how the uncertainty of the measurements is affected by a change of frames. That is, we determine what can an observer attached to a landmark frame deduce from the information transmitted by an observer attached to the robot frame. This SLAM system is ideally suited for the navigation of domestic robots such as autonomous lawn-mowers and vacuum cleaners.