| One of the most important prerequisites for the use of household robot is its indoor location.Visual sensors can provide abundant information and thus research on the location of vision-based household robot is of great significance.This thesis mainly focuses on the techniques about the indoor location of vision-based household robot,which carries an active cursor on its body and can be located by identifying the cursor.The main research contents are as follows:Firstly,the thesis constructs the mathematical model of camera based on the models of common cameras,combining the characteristics of the visual location of household robot.The visual model of camera imaging is established to make sure each area of the room to be covered visually by at least two cameras.Besides the thesis also study the common camera calibration method in the hope of specifying the exact way to mark cameras as well as algorithm of household robots based on monocular vision and two different location methods are designed according to the two different ways of target recognition.Secondly,the paper studies on the reconstruction method of indoor mapping and ultimately works out the method of identifying CAD to reconstruct indoor contour and displaying the way of interior contour reconstruction by recognizing CAD drawings.The realization of the two kinds of indoor positioning algorithm is studied in detail and the location results are shown on the 2D map.In addition,the ultrasonic sensor is studied as it can help achieve the obstacle avoidance function of the domestic robot in its walking process.Next,the prototype of household mobile robot is designed and the robot control system is built.Then the ultrasonic sensor is applied on the prototype so that the robot can avoid the obstacles.Lastly,the experiments on the prototype robot built before and the arranged cameras are carried on respectively,and the results to be gained for further comparative analysis. |
PDF Full Text Request