Underwater Target Tracking Algorithm Based On Interacting Multiple Mode

In recent years,the exploration of marine resources has been actively promoted.In the applications of underwater mineral resources mining,fishing,and underwater robot positioning etc.,the research of underwater target tracking is particularly important,which requires higher accuracy.During underwater target tracking systems,the target tracking algorithm is used to deal with measurement information including distance and angle so that the trajectory of underwater targets can be obtained.Considering current filtering and tracking algorithms,Kalman Filter(KF),Extended Kalman Filter(EKF)and Unscented Kalman Filter(UKF)are widely used in underwater target positioning.Kalman filter algorithm and its extended algorithms can match the non-motorized state of underwater targets.Once mobile state exists,the great deviation of actual target trajectory couldn't be avoided with the only use of Kalman filter algorithm to tracking and location.With input interaction,filtering,probabilities updating,output interaction and other steps,Interacting Multiple Model(IMM)algorithm which is in use of interactive multi-model technology can be adopted to match possible states of turning,acceleration,deceleration,etc.,and also,the defect that the traditional Kalman filtering algorithm only applies to non-motorized target tracking could be overcome.In this paper,an interactive multi-model Kalman filter algorithm is used to track underwater maneuvering target in the linear Gaussian dynamic systems.As for nonlinear Gaussian dynamic systems,combined with interactive multi-model algorithm,the extended Kalman filter algorithm and the unscented Kalman Filtering are adopted in turns for underwater target tracking.Finally,simulation results verify the effectiveness of the tracking algorithm.This paper studies mainly from the following three parts:1.Underwater target tracking algorithm based on IMMKF-3D:For linear Gaussian dynamic systems,the three-dimensional Kalman filter(KF-3D)algorithm owns high accuracy for motion estimation of targets,but only for the uniform linear state.As the underwater target is in a state ofmaneuver,the possible turning,acceleration,deceleration and other states can be disposed by using an interactive multi-model algorithm.Combining interactive multi-model algorithm with three-dimensional Kalman filter algorithm,a three-dimensional Kalman filter algorithm(IMMKF-3D)based on interactive multi-model which used for tracking 3D underwater target is proposed to realize the accurate tracking and positioning of underwater maneuvering targets.2.Underwater target tracking algorithm based on IMMEKF-3D:For nonlinear Gaussian dynamic systems,the extended Kalman filter algorithm converts it into an approximate linear Gaussian dynamic systems by use of Taylor series.Combining interactive multi-model algorithm with three-dimensional extended Kalman filter(EKF-3D)algorithm,a three-dimensional extended Kalman filter algorithm(IMMEKF-3D)based on interactive multi-model which used for tracking3 D underwater target is proposed to realize the accurate tracking and positioning of underwater maneuvering targets in nonlinear systems.3.Underwater target tracking algorithm based on IMMUKF-3D:Aiming at the problem that the nonlinear Gaussian dynamic systems linearization process is easy to diverge and has big error,with the combination of interactive multi-model algorithm and the three-dimensional unscented Kalman filter(UKF-3D)algorithm,a three-dimensional unscented Kalman filter(IMMUKF-3D)algorithm is proposed.The algorithm can reduce the errors brought by the linearization process and realize the accurate tracking and positioning of the three-dimensional underwater maneuvering target.The design scheme of this paper is in the background of the underwater three-dimensional space,and is verified by the MATLAB software platform for simulation.Experimental results show that the underwater target tracking algorithm based on interactive multi-model provides strong effectiveness for real-time positioning of the three-dimensional underwater maneuvering target.