Research On The SLAM Algorithm To The Application Of Underwater Vehicles Based On EKF

With the improvement of significance for the ocean strategy, the development of Autonomous Underwater Vehicle (AUV), which has became hot topic in research of robot technology during recent years, has attracted widespread attention from more and more countries. Navigation technology is the cardinal method to ensure AUV,s safety and autonomous voyage. High precision navigation system is also a prerequisite condition for successful process in ocean. Without dependence on priori information of underwater environment, Simultaneous Localization and Mapping (SLAM) method makes relationship between the collected essential signals via external sensors and position of geographical features to realize autonomous localization and map construction.In this paper, the expression methods for environmental map were discussed at the first.The feature map model was chosen to describe the environment in research. The feature’s position information was regarded as points to avoid the influence of their geometric shape.Then, conception and difficulties in SLAM method were introduced, including construction of mathematical model of system and the influence caused by data association process. To enhance the algorithm’s efficiency by sifting useful information from amount of collected Data, Nearest Neighbour Data Association (NNDA) was introduced and utilized to achieve the goal. However, in realistic condition, the model of AUV’s underwater movement was nonlinear, EKF (Extended Kalman Filter) algorithm was added to fulfill the design of whole filter process. Thus, the feaures’ information would be interrelated step by step to ensure the convergence of path pursuit and correlation between environmental features. Two voyage modes, including simple and complex movement, were discussed and simulated to verify the algorithm’s reliability and stability.To make ordinary application of algorithm, EKF-SLAM method was realized in 3D space. By making reasonable resolution of space movement, the proficiency of algorithm was enhanced and complicated calculation was avoid in data process. After establishing the new coordinate system and movement model, the prediction process and observation process were also improved. Finally, the simulation of algorithm application in realistic condition was fulfilled to further verify EKF-SLAM method’s feasibility. Research in this paper will offer reliable evidence of algorithm application in reality.