Key Technology Research On Simultaneous Localization And Mapping For Panoramic Mobile Robot

Simultaneous localization and mapping technology has focused attention of many authers in mobile robot navigation areas, and is considered the key technology to realize the autonomous mobile robot. Because the panoramic vision sensor has a360degree comprehensive perspective and information-rich so it is ideal for use in mobile robot navigation. The panoramic mobile robot SLAM technology combines the advantages of both, in this article.Firstly, the thesis describes the SLAM problem in a mathematical model, also the existing methods to solve the SLAM problems are summarized and reviewed, by comparing, the thesis choose the EKF algorithm, and was proved effective by simulation experiment. As the distortion of the panoramic image is large, the thesis propose to use a more robust feature extraction algorithm SIFT to deal this problem, and this algorithm has been proved stabley and robustly by the simulation experiment. The robust feature is useful for the three dimensional reconstruction afterly.Secondly, the thesis designs two methods for the calibration of the panoramic vision system, to make the calibration more pinpoint. After that, the thesis investigates the principle of formation of panoramic vision system, and based on this principle the thesis proposed a kind of theory and method to realize the three-dimensional reconstruction of the feature, and this process must assist with the data of the odometry. By the simulation experiment, this method of three-dimensional reconstruction algorithm has been proved higher accuracy, and can offer a high-accuracy map for the panoramic vision SLAM research.After that, the thesis designs a systematic structure of SLAM for mobile robot. It is based on existing panoramic mobile robot simultaneous localization and mapping methods. Meanwhile detail solutions and methods were given for the key technology and difficult point of every module. And then the thesis do a simulation experiment, in the experiment common vision SLAM and panoramic vision SLAM were simulated by setting different range of observation for the sensor. The simulation results show that a larger range of observation for sensor can help the robot to track the landmark more effectively, and the uncertainty of the SLAM system can converge more quickly. At last, the thesis do the SLAM experiment on the panoramic mobile robot, and the result prove the solution for the panoramic mobile robot SLAM problem is feasible and effective.Finally, a conclusion of panoramic mobile robot SLAM problem is given, and the expectation is made as well.