A visual SLAM system on mobile robot supporting localization services to visually impaired people

摘要: This paper describes a Visual SLAM system developed on mobile robot in order to support localization services visually impaired people. The proposed aims provide small or mid-scale environments such as inside building campus of school where conventional positioning data GPS, WIFI signals are often not available. Toward this end, we adapt and improve existing vision-based techniques handle issues the indoor environments. We firstly design an image acquisition collect visual data. On one hand, robust odometry method is adjusted precisely create routes environment. other utilize Fast-Appearance Based Mapping algorithm that may be most successful for matching places large scenarios. In better estimate robot’s location, Kalman Filter combines results current observation estimation states based its kinematic model. experimental confirmed feasible navigate people