Research On Simultaneous Localization And Mapping And Path Planning Of Underwater Vehicle

Autonomous Underwater Vehicles(AUV)play an important role in the exploration and development of deep-sea resources.However,if an AUV wants to move autonomously,it is always inseparable from positioning,mapping,and navigation,and simultaneous localization and mapping(SLAM)and path planning are the key points.SLAM algorithms are mainly divided into two categories: filtering and smoothing.Compared with the limitations of filtering algorithms,smoothing algorithms have more advantages,and the incremental smoothing and mapping(i SAM)is the dominant smoothing algorithm.The ant colony algorithm is an intelligent optimization algorithm.Compared with other path planning algorithms,the ant colony algorithm has the advantages of good feedback information,strong robustness,and a good distributed computing effect and has achieved good results in the application of path planning for underwater vehicles.In order to improve the efficiency and accuracy of the incremental smoothing algorithm,the motion error learning model of the AUV is established using Gaussian process regression,and the interference of the underwater environment on AUV motion is fitted into the error learning model,which is combined with the least squares problem model of smoothing SLAM so as to reduce the interference of ocean currents and other factors on the AUV motion model and improve the accuracy of the algorithm.At the same time,it is observed that QR factorization induces more non-zero elements than Cholesky factorization in the calculation process,thus slowing down the mathematical operation of the sparse matrix involved in the calculation.Therefore,Cholesky decomposition is used to replace the QR decomposition of the original i SAM algorithm,which improves the efficiency of the algorithm.Finally,five principles of flexible reordering are proposed to improve the periodic reordering of i SAM and improve the efficiency of the algorithm.Aiming at the problems of low efficiency of path search,easy falling into local optimum,and slow convergence speed when ant colony algorithms are applied to AUV path planning,a combined AUV path planning algorithm is proposed.In view of the fact that the probability selection of the traditional ant colony algorithm is not always guaranteed to be the optimal solution,and sometimes it even falls into the local optimum at the initial stage of the algorithm,a combination operation sp is designed to improve the state transition probability in the traditional ant colony algorithm so that the ant colony can better select the next grid.Aiming at the problem of limited heuristic search and poor convergence speed of the traditional ant colony algorithm,new heuristic information is adopted according to the principle of infinite step size,which expands the field of vision and improves the visual accuracy.At the same time,the improved algorithm adopts new pheromone updating rules and dynamically adjusts the pheromone volatilization factor,which accelerates the convergence speed and expands the search space.Finally,a rollback strategy is designed to solve the problem of an AUV falling into underwater obstacles such as coral.