Control System For A Vision-Based Wheeled Mobile Robot Based On SOPC

Vision information processing has been challenged by the problem of real time capability for long. In the ecological environment involved with the technologies mainly as Integrated circuit and development tools, hackneyed Vision-Based Wheeled Mobile Robots have too simple functions and poor performance to meet the application's needs. Directed against the status quo, an investigation is implemented on Software-and-hardware-coordinate image processing system based on SOPC (System On a Programmable Chip) aim at real-time capability improving. There is an involved job of designing a vision-based wheeled mobile robot based on SOPC.This paper presents an analysis on the features of vision information processing, as well as, according to the features, the architecture for Software-and-hardware-coordinate image processing system based on SOPC. In the architecture, a histograph stat. circuit based on PLB (Processor Local Bus) and a neighborhood exacting circuit is designed, and 70% work for the video capture system base on the PLB bus is achieved. The neighborhood exacting circuit is verified with the experiment of edge detection and space filter processing, and the video capture system is verified with on-chip logic analyzer and memory viewer, the whole architecture for Software-and-hardware-coordinate image processing system is evaluated by Software-and-hardware-coordinate Otsu segmentation.Additionally, this paper presents the design of a vision-based wheeled mobile robot named VNTES-1. Involved designs include a precise Mech-electronic controller based on CPLD (Complex Programmable Logic Device), Zigbee wireless network for mobile communication, embedded software for image processing and control arithmetic implementation, dynamic path identifying arithmetic as well as multi-point preview arithmetic for path tracking. These works are verified reasonable by the on-road test with VNTES-1. Besides, a polynomial fitting arithmetic for is used for depth of field extraction while points of intersection filter and weighted-least-squares arithmetic is used for vanished point seeking.The conclusion is that the Software-and-hardware-coordinate image processing based on SOPC can distinctly improve the real-time capability of vision information processing. Furthermore, the control system for a vision-based wheeled mobile robot based on SOPC is able to achieve a serial of task well, such as vision information picking-up, move decision-making, mech-electronic control and mobile Communication as well. The system has quite well a prospect for application.