Research On Target Estimation And Trackingin AUV Path Planning

The importance of Autonomous Underwater Vehicle(Autonomous Underwater Vehicle,AUV)in engineering applications,marine defense and so on,making the development of related technology has become a hot research during the countries in the world.Path planning technology is an affordable prerequisite for AUV completing the task successfully.To study the path planning,the priorityis to make a valid estimate and real-time tracking of the underwater target,and we can see how to achieve tracking the goal of underwater environment is very important.I have deeply studied the target motion coordinate system selection,the state of motion modeling,and objective estimation problem in this paper,andanalyze the application of the nonlinear filtering method in target tracking.The details are as follows.When building the model for the state of the underwater target,I select the space Cartesiancoordinate system and spherical coordinates whose relative motion are the most useful.At the same time the expression of the information of the location,speed and acceleration of the target during the two coordinates are being showed,and I have a detailed derivation for the conversion relationship in a different coordinate system of theseinformation.Then building the target motion model and proposing Target motion modeling based on linear systems and nonlinear systems,the analysis gets the state equation and scope of different models.It should choose a singer model which owns a better description of the target linear motion,and use Kalman filter for target motion tracking.So I have studied the content of the discrete Kalman filter algorithm implementation steps,and I get the analysis of the algorithm error--Variance error method,curbed the filtering divergence with the recession mnemonics.I made a simulation experiment achieve tracking the target motion by the filtering algorithms and analyzed the results of the experiments.According to the necessity of the nonlinear target motion tracking,three kinds of nonlinear Kalman filter are studied in this paper.One is based on the Extended Kalman filter function approximation,the other is based on Unscented Kalman Filter deterministic sampling,and the last is Cubature Kalman Filters which is based on Spherical-Radial cubature rule.The principle and implementation process of the three kinds of algorithms are given,and I have analyzed the advantages and disadvantages of the algorithms.Finally,I applied the algorithm to three-dimensional target tracking simulation experiments,analyzed the performance of different algorithms.Of the three nonlinear target tracking algorithms,the best one is CKF while the lowest precision is EKF,UKF is the center.But if the nonlinear maneuvering target movement of the simulation is not strong,the three algorithms tracking error are within the allowable range.