The Research Of Automatic Guided Vehicle Based On Visual Navigation And Tracking Control

Automatic Guided Vehicle (AGV) has the ability to drive under unstructured circumstances automatically, and it can be applied in many fields. Navigation is a key technique among the researches of AGV. Based on mark line recognition navigation, it becomes a main research direction because it has more merits than other visual navigation methods, such as easy setting up and changing of mark line, low technical cost and so on. Meanwhile, it can quickly process the images and has a good performance of real-time control. Aimed at mark line navigation, this dissertation takes car-like AGV as the research objective, suggests a navigation route recognition algorithm based on linear model, which resolves the problems such as noise elimination, the edge of the mark line obtaining and the location of the AGV during the process of image processing. The simulation results show that the algorithm can satisfy the requirements of navigation control system of the AGV. Tracking control is another key technique among the researches of AGV. Therefore, it is very important to design a control system which has the following characteristics: small tracking error, quick dynamic response, adaptability to complicated circumstances and robustness and. so on. When the AGV is running, if the steering angle is smaller, the optimal controller has good performance, but it cannot be adapted to steering control in case of larger angle. Fuzzy controller is suitable to this case, but the control scheme cannot cover all situations. Based on this, a control system combining fuzzy and optimal control methods is also presented. The results of simulation show that the control system could make the AGV have the adaptive capability to track various mark lines, have the strong capability against noise disturbance and have good performance on control effects. In order to verify the simulation results, an AGV experimental platform with visual navigation is implemented, and the software and hardware of the real-time control are developed in order to realize the tracking control of the AGV. The experiment results show that the real-time control system of the AGV can efficiently perform the real-time control task, and possess the characteristics with little tracking error and quick dynamic response.