Performance analysis of Bayesian target tracking method

One of the most important parts in target tracking is the filtering algorithm.The typical Kalman filter can get the recursive minimum mean-square estimation under the linear and white Gaussian noise circumstance.Compared to the Kaman filter,the α-β-γ filter reduces the computation complexity and can be easily implemented in engineering applications.Based on the principle of Bayesian filtering method,this paper analyzes the estimation accuracies of Kalman filter and α-β-γ filter,and gives the applicative conditions of the two different methods.The fundamental theories and filtering algorithm implementations of Kalman filter and α-β-γ filter are studied.Simulations of tracking the targets with uniform velocity and acceleration in the straight line are presented.The simulation results show that the Kalman filter,α-β filter and α-β-γ filter are all with better tracking capabilities and good real-time performances in linear and Gaussian environment.