Applying Grey-Fuzzy-Fuzzy Rule to Machine Vision Based Unmanned Automatic Vehicles

Abstract

The unmanned automatic vehicles (UAV) are important elements in industrial application, such as unmanned carriers in warehouse. With the great progress in the intelligent transportation systems, the UAVs are often applied to drive automatically. The mechanical vision replaces the human’s eyes to be the feedback component. By the image processing, the necessary information can be figured out. However, due to the complicate calculation of image process, it will take a long calculation time from fetching image to making image information. Hence, a novel control rule that is the Grey-Fuzzy-Fuzzy controller is proposed in this paper for vision guided control of UAVs. The novel rule consists of the base-layer and upper-layer fuzzy controllers. The base-layer fuzzy controller is the classical fuzzy controller using the position error and the error difference to be the input variables. The Grey prediction is used to estimate the position error at the time of making image information. Then, the upper-layer fuzzy controller uses the two input variables, which are the estimated position error and the estimated error difference, to correct the base-layer fuzzy controller’s output signal. Finally, an experimental UAV is developed to examine the potential of the proposed control scheme. The vision-based experiment of automatically driving of lane-following is carried out in this paper. The experimental results show the proposed controller will eliminate the swinging phenomenon and increase the accuracy while tracking. The experimental results also show the practical capability of the Grey-Fuzzy-Fuzzy controller.