Research On Mobile Robot Based On Kinect Depth Information

With the rapid development of the technology, there is generating new technology in the field of mobile robot navigation, mobile robot using monocular camera, ultrasonic, proximity switches, laser sensors, is not flexible enough in navigation and obstacle avoidance and expensive. This article uses Microsoft’s new release of the product as the core Kinect sensor, and uses Turtlebot robots as an experimental platform to study the localization of mobile robots in unknown environment, creating maps, obstacle avoidance and path planning related content.This paper described the research status of the key technologies of mobile robot navigation. Summarized and analyzed the existing problems. Divided the navigation technologies into simultaneous localization with mapping and obstacle avoidance with path-planning, the contents and results are as follows:Firstly, Described the technical principles of Kinect sensor and the extraction process of the point cloud data outputting into an image. Elaborated the theory of several common classical algorithms during the process of SLAM, comparing and analyzing the advantages of different kinds of algorithm. Chosen Gmapping feature pack as experimental support of this paper and described several key steps of Gmapping in detail; Secondly, Study on the obstacle avoidance mechanisms of mobile robot, including the detection of obstacles and local map updates, so that the construction drawing stage which can be carried out manually and automatic control.The key content of path planning is described. The theory steps of A* algorithm, depth-first algorithm, breadth-first algorithm and Dijkstra algorithm are elaborated. Last, the process of path planning and avoiding and the process of nodes searching are simulated. The detection and matching of specific nodes are verified and analyzed based on turtlebot and Ubuntu system, while location, drawing and path planning and avoiding are also verified.As for our laboratory, the research results of this paper, which include mapping navigation in an unknowing environment and application researching under Ubuntu Linux system, are belong to a new field and may be the basis of later researching.