Research For Mobile Robot Self-localization And Navigation Mapping

In the indoor environment, the mobile robot self-localization and navigation is one of hotfields of research. The mobile robot can identify the character information, which is the basisof mobile robot can realize self-navigation. And an artificial landmark in designing, due to itsshape, size and color can be set in advance, so, when a mobile robot carries a navigation task,artificial landmark is the first choice. How to design a new type artificial landmark and themobile robot can identify it quickly and accurately is the focus of research.. The traditionalself-localization methods is as follows, dead reckoning method, method of mode matchingand combination method, etc, yet, each method has its shortcomings and insufficiency.Simultaneous Localization and Mapping(SLAM) can build a navigation map in the process ofself-localization, which provides a new solution to the mobile robot self-localization andnavigation, besides, vision sensors are the characters of informative and easy equipment,therefore, the vision SLAM is the development trend in mobile robot self-localization andnavigation.This topic takes the intelligent mobile robot as the platform for research, carry out thefollowing works, such as new artificial landmark design, self-localization based on newartificial landmark, template tracking based on natural landmark and building navigationmapping based on mix landmarks, the specifically research works are as follows:Firstly, by using the principle of BCH code, we design a new type of artificial landmark,this design is not only easy to make and rich information, but also has a certain errorcorrection function. In the editing interface, as long as input relevant information, you cangenerate landmark design, in addition, this design can be printed to use directly. In theexperiment, the mobile robot can not only recognize the signs of information quickly, but alsowith the character is the low error rate, and good robust.Secondly, in the design of artificial landmark, choose the pentagon as landmark outline,so, the recognition algorithm adopts Canny operator. In the basis of new artificial landmarkcan be identified accurately, mobile robot realizes self-localization by use OI algorithm. In thestate of static and motion, the landmarks are posted on the ceiling and vertical planerespectively, we obtain the position error curve and angle error curve respectively bysimulation.Third, take use of feature invariant factors, then, extract feature points of naturallandmark, filter out the interference point and determine the template contour point, afterthese steps, the system generates a template of landmark information, on this basis, we havecompleted the experiments of template-tracking in basis of natural landmark, and set therange of mobile robot can track.Finally, in the basis of natural landmark and artificial landmark, by use visual SLAMprinciples: in the motion, mobile robot can build a navigation map, we construct a navigationsystem, and give the software structure of this system, then, propose three ways to build the map. In indoor environments, artificial landmarks are affixed to the ceiling, according to thethese landmark distribution to build a topological map, by use this map information, mobilerobot is able to complete self-navigation task, in the vertical direction, make use oftemplate-tracking in basis of natural landmark, the mobile robot performance visual servo.