Research Of Robot Vision Navigation Based On FPGA

The navigation technology of robot is the important subject in the mobile robot research field.With the development of sensor technology, it emerges related fields:vision navigation, electromagnetism navigation, photoelectric navigation, inertia navigation, olfaction navigation and other different kinds of mode.But among them,Vision navigation is the backbone in the practical application.It is essential in research to improve the real-time capability of image processing and the accuracy of path recognition.Because of the low processing speed and high cost of hardware,the traditional system which consisted by embedded MCU has restricted the development of vision navigation. With the development of FPGA (field programmable gate array) technology, its SOPC system has the advantages of high integration, small volume, low power consumption, and easy to change and upgrade.In taking these advantages of SOPC system,it will greatly promote the real-time image processing and its entire reliability.This paper is focusing on the study of the system, using linear CCD to gather the image, which controlled by the FPGA, driven by the stepper motor. Replacing the traditional image sensor, the CCD not only eliminates the image distortion, but also solves the problem of traffic.Through FIFO buffer memory image data storage, using the FPGA realizing the image preprocessing based on the hardware structure and Avalon, by bus NIOSII embedded processors communication software realization pipeline image processing mode, this structure can greatly improve the real-time visual navigation. In the path of the calculation process of identifying motion control based on the introduction of acceleration sensor, navigation path of inertial navigation system is smoother. Using the FPGA embedded processors NIOSII soft control drive stepping motor navigation. In the design of hardware, it combines the Verilog HDL and the principle diagram language, adopting the way of refining from top down.This design realizes the integration of the CCD control, path linear CCD image preprocessing, binary NIOSII embedded processor, stepping motor control into a FPGA. Compare with the traditional mode of modular design, it has the advantages of high integrity, good real-time, and high reliability. Combining with the QuartusII simulation software validation it is more convenient to make its verification without any changes of the hardware so as to shorten the development cycle to improve efficiency.