The Study, Based On Pmac's Vision Navigation

Detailed studies are made on the AGV motion controller based on PMAC (Programmable Multi-Axes Controller) and the color vision-based navigation of the AGV in this paper. These studies provide evidences for improving the precision and the reliability of the AGV.In this thesis the QDU- I vision-based AGV is successfully designed and developed, and a new type of AGV motion control system based on PMAC is designed too. PCOMM32 is used to redesign the PMAC under the environment of VC++ 6.0. According to the practical requirements, the software system and design environment of the QDU-I are decided. Multithread technique is adopted to the design of the software system and the user interface in this thesis. On the basis of the analysis of the method and theory of camera calibration, a camera calibration method based on polynomial fit for the vision system is put forward in this paper, and a following test proves its efficiency. A new, simple and quick filter approach is introduced to recognize the navigation landmarks of the AGV. This approach successfully resolves the real-time problem of the color image processing, using the equality among the value of R, G and B of the gray scale. This thesis proposes a least square method to fitting the landmark line. According to the possibility of the damage of the landmarks and invalidation of image processing, a amelioration is made to the above method. In order to meet the practical requirements of the AGV, three special signs: turn left sign, turn right sign and stop sign are designed in this paper. At the same time, a fuzzy control program is designed for the QDU-I , and prove the feasibility of this fuzzy control system through a path following test.