Research Of Simultaneous Localization And Mapping Using On AUV

The work of this thesis is focused on the Simultaneous Localization And Mapping (SLAM) technology on the Autonomous Underwater Vehicles (AUV) application. It is part of the "The Distributed Intelligent Control of Multiple Autonomous Underwater Vehicles Project" sponsored by CSTIND (Commission of Science Technology and Industry for National Defense).A SLAM technology based on Doppler Velocity Log (DVL), Fiber Gyro and Forward Looking Image Sonar equipped by a typical AUV is presented here. Firstly, we give a summary of current development and the state of art of the worldwide AUVs research and construction. Then in the following section the theory of the SLAM technology and the methods to describe the environment in which an AUV will navigate are introduced, and we adopt the segment-based environment representation. About the look-forward sonar data, using The Extended Kalman Filter (EKF) and some segment fitting technologies are presented in succession, which are used to detect the edges in the Forward Looking Image Sonar data processing. The fourth section gives a description of the localization problem as the matching between observed features and map features, using the IWL algorithm. The formal description of the SLAM technology and its application in AUV controller are discussed in the last chapter, and the experimental results, both of tank test and of simulation, demonstrate its validity and advantages.The results of this paper demonstrated the significance and advantages of the SLAM technology in the application of the AUV navigation and map building.