Vision Guidence Research Of AGV

AGV is the short of automatic guided vehicle. It is an essential tool of modern logisticsand industrial manufacture. Using AGV improves the logistics speed, reduces costs andshortens the production cycle. In a word, the biggest work of AGV is to extremely increaseproductivity. Therefore, it plays an irreplaceable role in creating higher value. And theproposed project is based on this background. The core technology of AGV is the way ofnavigation. Linux-based embedded visual navigation system has very excellent scalability andresearch value with the big development of machine vision and embedded technology.The AGV visual navigation system design involves two aspects: image capturing&processing and controller. In software section, we complete the image capturing&processingand controller design.After the image capturing and dealing with the noise, we design a fast dynamicbinarization technical to overcome the defect that it takes too much time to calculate toprocess image. And the defect does not meet the requirement of real-time performance ofembedded system. Designing a simple and fast technical extracts eigenvalues. We design aCMOS camera calibration algorithm to get the accurate controller input.To the controller part, we design a approach tracking algorithm for the fuzzy controllerbased on multi-pre-tracing point. In order to improve the real-time performance, this projectuse Matlab to get the offline fuzzy control query tables. Although this method increases thecomplexity of debugging, the real-time performance of system has been greatly improved.The system runs on Linux operating system based on ARM11S3C6410developmentboard. The design of image processing part uses Codeblocks IDE, gcc compiler and gdbdebugger. The design of controller part is simulated by Matlab.Experimental results show that the system design of this project can track the path quickly,accurately and steadily, the real-time performance and robustness of the system meet thedesign requirements. The project has certain reference value to AGV visual navigation andbroad application prospects.