Research On Target Tracking Based On Cubature Kalman Filtering

The wireless sensor network is one of the important three-dimensional network research technologies because of its sensing and wireless communication functions.The sensor network is composed of multiple nodes with self-organization capabilities.The wireless sensor network is currently a popular research field in many underwater applications,such as military prevention monitoring and seabed exploration.Target tracking is an important technology of three-dimensional network development research,and the current state of underwater target movement is estimated using data information provided by multiple sensor nodes.In the target tracking system,use filtering algorithm to estimate target motion state,and the related uncertainty is eliminated according to the filtering algorithm to realize the system tracking the target.Therefore,this thesis mainly studies the target tracking based on cubature Kalman filtering,and improves the tracking efficiency by improving filtering algorithm.This thesis studies the problem of target tracking based on wireless network.Firstly,the basic characteristics of the wireless networks,and the target tracking system model and three common target motion models are analyzed,including the constant velocity(CV)model,constant acceleration(CA)model and constant turning(CT)model.The core technology of the target tracking algorithm is the filtering algorithm.The several common filtering algorithms in the tracking system are introduced,including Kalman filtering(KF)algorithm,extended Kalman filtering(EKF)algorithm,unscented Kalman filtering(UKF)algorithm and cubature Kalman filtering(CKF)algorithm.Compare and analyze these several nonlinear filtering algorithms,and the cubature Kalman filter is more suitable for practical target tracking applications.Most of practical applications are nonlinear systems,and the nonlinear filtering is a hot research in target tracking.The linearization Kalman filtering has many flaws.Therefore,an improved cubature Kalman filtering algorithm(ICKF)is proposed for target tracking.There is uncertainty in the target movement.According to the strong tracking principle,an adaptive forgetting factor is given into the CKF algorithm to directly modify the error covariance to reduce the impact of uncertainties.Then,because of variable target motion,interactive multi-model(IMM)technology is introduced to solve the problem of single target motion.Compared with other filtering algorithms,the IMMICKF algorithm considers the improved cubature Kalman filtering algorithm and the IMM algorithm,which can effectively solve the nonlinear tracking problem and get better estimation results.Finally,the marine environment is a typical three-dimensional network.Due to the complexity of the underwater,the batteries carried by the sensor nodes in the wireless network have limited energy and cannot be replaced.Therefore,from the perspective of the limited energy of underwater sensor nodes,an energy-efficient CKF algorithm is proposed,which comprehensively considers the contribution value of sensor node transmission data and residual energy information.Balance the wireless network energy usage,and use the non-linear cubature Kalman filtering to achieve underwater target tracking.The proposed algorithm improves the overall tracking performance of the system.