| The visual navigation technology for mobile robot is the research hotspot in therobotics field. How to improve the real-time and flexibility is urgent issue for the visualnavigation. In the paper, aiming to study the visual navigation based on landmarkrecognition, the hardware and software based on the embedded system are designed, inaddition, a digital signpost recognition algorithm and a corresponding navigation strategyare proposed. As a result, the path planning for mobile robot is realized under theguidance of digital signpost. This thesis include the followings:(1) The hardware platform of the mobile robot is built. The embedded processorS3C2410 is taken as the data processing and control centre, and the OV7620 camera isused for image acquisition and the motor driver module is designed based on the L298chip.(2) The software system is designed based on the above hardware platform. In thesystem, the Linux operating system is cut and ported to the core board. The device driversof ultrasonic module, camera and DC motor are developed with the software GCCcompiler. In addition, The human-computer interface was designed with the software QT.(3) The path planning method for mobile robot is researched based on digital signpostrecognition. A new recognition algorithm is designed based on the threading method andprogrammed with C language. Then, in order to improve the traditional chessboardmethod based on local map building, a unit-based path planning strategy is proposed,which realizes the path planning for mobile robot under the guidance of digital signpost.(4) Finally, experimental research which aims to realize the path planning based onchessboard method and the recognition of the digital signpost based on the threadingmethod is done to validate the navigation of the robot. The results show that the pathplanning of the mobile robot guided by digital sign is effective and reliable. |
PDF Full Text Request