High Precision Localization Integrating Environment Mapping And Sparse Artificial Landmarks

The rapid development of mobile robot technology brings some critical problems to be solved urgently.It is a research hotspot to improve SLAM(Simultaneous Localization and Mapping)accuracy while ensuing consistent performance in a complex and dynamic environment under the premise of meeting the time-space complexity requirement as well as the hardware.An extended SLAM method is developed in this thesis towards this objective.While integrating landmarks with the natural environment localization module,mobile robots can not only improve the localization accuracy,but also make it more reliable based on the measurements about the artificial landmarks.On the other hand,when the natural environment localization result is added into the artificial landmark based localization module,it becomes easier to match some landmarks during motion,and also to can eliminate the cumulative error due to inertial navigation.The main work of this thesis is summarized as follows:1.A localization algorithm is proposed based on the relative position measureme nt of sparse artificial landmarks.The algorithm is divided into static and dynamic localiza-tion modes.The static mode is responsible for the acquisition of the initial pose of the mobile robot and the dynamic localization mode is to estimate the pose of the mobile robot after the initial pose acquisition.The algorithm will not be influenced by the dynamic environment,which is validated in a real mobile robot,platform.2.A natural environment based localization algorithm is proposed integrating environment mapping and sparse artificial landmarks.The algorithm adds the extra constraints generated by the landmark based localization module to the cost function of loop closure.Thus,the algorithm is adaptive to dynamically changing environment in terms of localization accuracy and reliability.