Research On Radar Target Tracking Algorithm Based On Kalman Filtering

Radar target tracking plays an more and more important role in the area of detection and early warning system,it is also an important study topic of computer fusion.In many fields like air defense warning,operation decision,mineral excavation,,traffic control,target tracking is a basic theoretical problem.The design of the radar tracking filtering algorithm is the most important part in target tracking system.The Kalman filter,which is based on state space model in the time domain,overcomes the shortcomings of the Wiener filter which can only be worked in the frequency domain and cannot estimate the multi-variable,non-stationary random process.Among the various approaches to estimate the state of target tracking problem,Kalman filter is drawing more and more attention with outstanding performance and strong noise suppression capability.The thesis first starts from the target state space model,analyze the theory and characteristics of various target kinematics model.the thesis discusses the different two coordinate systems and the transformation between rectangular coordinate system and polar coordinate system.In the next,the thesis present several filtering approaches for target tracking problem,in consideration of that the actual radar system contains many nonlinear variables,the Extended Kalman filter(EKF)and Unscented Kalman filter(UKF)are discussed primarily,the MATLAB experiments have shown that the Unscented Kalman filter is more excellent in target tracking process than the Extended Kalman filter.At last,the thesis studies the merits of Interacting Multiple Model(IMM)algorithm basing on Kalman theory for the maneuvering target.And an attempt by combine Unscented Kalman Filter and Interacting Multiple Model algorithm is proposed.Through the MATLAB experiment we can find that the attempt has a better performance in practical radar target tracking problem.