Semi-guided navigation of AGV through iterative learning

In this paper, the authors aim at realizing an accurate navigation system of automated guided vehicles (AGV). The authors propose a way of estimating positioning error with magnetic tape, which is widely used in a factory as an external sensor. However, flexibility for path relocation is insufficient, because, in general, the tape should be laid down on the floor from a start point to a goal point so that AGV can reach their target. To overcome this inefficiency, the authors firstly propose a semi-guided navigation methodology by means of two kinds of magnetic tapes based on an error analysis. The semi-guided navigation means that magnetic tapes are only placed at the start and the goal points individually. Therefore, this system enables us to remove most of the magnetic tape. Moreover, the authors attempt a fixed model learning to prevent stationary error while AGV run iteratively. Finally, the authors carry out experiments to evaluate and verify the efficiency of the proposed method