Target Tracking With Mixture Filter Gaussian

Radar target tracking, the state of the general objectives established in theCartesian coordinate system, and the target measurement information given in sphericalcoordinates or polar coordinates. No noise spherical coordinates unambiguous Dopplermeasurement information used in the Cartesian coordinate system is a nonlinear radartarget tracking.To solve this problem this paper describes the most common target trackingmethods, such as Kalman filter, unscented Kalman filter. Secondly, based on unscentedKalman filter this paper proposes a Gaussian sum unscented Kalman filter. Gaussiansum represents the probability density function of Doppler measurements. Estimationbased on Gaussian and the unscented Kalman filter approach to solve the ambiguousstate of Doppler measurements. The simulation results show the effectiveness of thisapproach. This method is unscented Kalman filter and Kalman filter to convertmeasurements showed a relatively higher accuracy in terms of position and velocityerrors.Finally, focused on a bit of Gaussian and parameter estimation problem. BecauseGauss and is the basis of our above-mentioned filtering algorithm. This paper alsopresents two of the most effective methods EM algorithm and particle swarm (PSO)algorithm. EM algorithm is also known as the expectation maximization algorithm,mainly through the steps and expectations constantly updated expectation maximizationstep estimates obtained, we believe that until you reach a reasonable result. ParticleSwarm method to estimate parameters of Gaussian and the group is mainly used tocomplete the search feature. Completed within a specific range of optimal estimationsearch other individuals through mutual collaboration between individuals and groups...