Design and Implementation of an Advanced Visual Guided Mobile Robot: ShinPu #1

In this paper we present the development and the functionality of an advanced visual guided mobile robot, namely "ShinPu #1", and demonstrated its two kinds of tracking applications, namely point tracking and path tracking applications. For point tracking applications, ShinPu #1 contains a visual servo system based on the sliding mode control technique such that it can automatically detect the shape features of objects and track them in real time even subject to partial occlusion and uncertainty of object movements. For path tracking applications, ShinPu #1 contains an efficient lane-marker tracking system, incorporated the inverse perspective mapping, the randomized Hough transform method and the least-square line fitting technique, to measure relative locations of the lane-marker at a look-ahead distance. In addition, by utilizing the feedback linearization technique, ShinPu #1 is able to navigate along a lane-marker and maintain a safe lateral offset to the lane-marker