| The investigation of robot self-localizationis very important to develop autonomousmobile robot and is a prerequisite for robot navigation and other intelligent behaviors.In this thesis, based on the computer vision, a study has been conducted based onpassive infrared landmark by studying the homographic positioning algorithm.Complete applicable indoor localization system for mobile robot with a core partdeveloped by using embedded processor was constructed. The main contents of thisthesis are as follows:1. A new kind of passive infrared landmark is designed which is suitable forself-localization of mobile robot moving inlarge range in indoor circumstance. It is ofeasy to be detected and to be recognized, of simple to location calculation, and ofconvenience to be manufactured.2. By analyzing the causes producing the localization deviation, a new algorithmmethod based on vertical assumption was put forward to compensate the tilt ofcamera. It is simple to calculate and can implement initial calibration of localizationmodulus rapidly. Localization deviation induced by actual installation error can becompensated by using optimized localization algorithm based on homography matrixand initial localization method of camera installation location. This algorithm is easyto realize in embedded processor.3. According to experimental analyze and localization results, it is shown that robotlocalization modulus based on embedded system combined with infrared landmarkcan achieve the localization function of robots. Furthermore, developed algorithm canbe used for robot localization with the satisfaction of localization precise. Thislocalization modulus can be used in indoor circumstance and hard to be disturbed bythe light from the outside. The usage of this modulus can reduce the cost of suchrobots using localization sensors. |
PDF Full Text Request