Study On Mobile Robot Navigation System Based On Global Vision

In the competition of Asia-Pacific Robot, whether the robot can travel fast and accurately is the key to determine to win or lose the test. So far, the navigation mode which most robots have adopted such as the transmission line and the coded disk gyroscope, is easy to be influenced by the factors especially like the surrounding environment of the playing field and the initial position of robot. These reasons will make the robot travel far away from the original trajectory and lead to the failure. In order to solve this problem, this paper proposes a way for navigation system of the robot which is designed based on global vision.This visual navigation system collects the venue environmental information by using the camera which is fixed at the top of venue, through processing and analyzing the collected images, the map model of the whole venue can be constructed. Then to calibrate the robot in the map model which is constructed before, so the real position of the robot in the venue can be located. Then, according to the actual requirements, do the robot path planning to make the robot to finish the assigned mission better.This paper mainly includes:the robot positioning and path planning.First, according to indoor environmental characteristics of the ground, proposed a positioning method by building a model of the environment map. This method used semi-thresholding and connected region determining to do the robot image recognition, determines the robot's position in the image coordinate system by the centroid calculation, through the inverse perspective transformation to build the connection of the world coordinate system and the image coordinate system, then according to mapping to realize the robot positioning.Then, according to the analysis of the current typical path planning algorithm and the specific experimental environment, the paper used a path planning algorithm which was combined by a grid method and the heuristic search algorithm, elaborated the principle of the algorithm in detail. In indoor environment, the paper did the path planning experiments and analyzed the results. The experiments showed that the visual navigation system met the design requirements, and the robot could realize self-contained navigation in the indoor environment.