Person Following using a Vision-Guided Mobile Robot
Person following with a mobile robot is one of the important research topics in the machine vision area. It is believed that mobile robots of the future can follow people intelligently in the crowded area. We have developed a person following mobile robot with maintaining a constant distance from a person using two cameras each of which has its own pan/tilt unit (PTU). Since our approach uses image-based PTU control and simple geometries for estimating a distance-to-target, it doesn't require calibrations for intrinsic parameters of cameras and stereo camera calibration. Instead of those, we use correspondences between pixel displacements in the image planes and pan/tilt angles of cameras, and also use modified laws of sine. For the target tracking algorithm, color histogram method is used. Our approach can be applicable to staff-following part delivery mobile robots in the manufacturing areas.
The technical details of this project will be shown in the publications soon, and the demo in the video shows how it works.
    •    Hyukseong Kwon
    •    Youngrock Yoon
    •    Jae Byung Park
Learning phase of the color histogram of the target
    [mpeg, 4.7MB]
Person following demo
H. Kwon, Y. Yoon, J. B. Park, and A. C. Kak, "Person Tracking with a Mobile Robot using Two Uncalibrated Independently Moving Cameras," Proceedings of IEEE International Conference on Robotics and Automation, Barcelona, Spain, April 2005 [ pdf, 642KB
    •    FINALE: Fast Vision-Guided Mobile Robot Navigation using Model-Driven Reasoning and Prediction of Uncertainty
    •    FUZZY-NAV: A Vision-based Robot Navigation Architecture using Fuzzy Inference for Uncertainty-Reasoning