| In recent years,with the advancement of technology and the development of the economic level,the demand for mobile robots in various industries is increasing day by day.As a key technology in mobile robot technology,mobile robot localization technology is also a difficult point in research,which has attracted wide attention of scholars at home and abroad.Aiming at the demand of mobile robots localization with high accuracy and no prior information,this paper uses cheaper 2D LIDAR as environment sensing sensor and adopts the indistinguishable artificial reflectors as the features to locate the mobile robot.The mobile robot is also located with reflectors and odometer by the extended Kalman filter algorithm to increase the accuracy and robustness of the localization.The main works are as follows:Firstly,the coordinate system of the mobile robot is defined.The motion model of mobile robot based on odometer and the sensor model based on LIDAR are also deduced.By comparing the map representation commonly used in mobile robot localization,the feature map is selected in this paper.Secondly,according to the characteristics of the reflectors,the reflectors existing in the environment are extracted and the centers of them are fitted by trigonometric function method.A triangle matching algorithm based on the triangle formed by the reflectors is proposed,and its flow diagram is elaborated.The principle of trilateration used in the global localization algorithm is described mathematically,and its solution is given in general.After that,the odometer-assisted method is used to eliminate the influence of LIDAR motion distortion on the localization accuracy of mobile robots.The odometer prediction model and LIDAR observation model are established,and the extended Kalman filter algorithm is used for fusion localization.The kidnapped mobile robot problem is analyzed,and solved by combining the global localization algorithm and the extended Kalman filter localization algorithm.Finally,the accuracy and robustness of the proposed algorithm are verified.By setting up the experimental platform and building the reflectors map,it is verified by experiments that the proposed global localization algorithm has the same localization accuracy as the commercial SICK Nav350 LIDAR.The experiment results also show that the performance of the extended Kalman filter localization algorithm is better than the global localization algorithm and the kidnapped mobile robot problem can be satisfactorily solved by combining them. |
