A Bayesian filter target tracking algorithm

Abstract

The invention discloses a Bayesian filter target tracking algorithm. The method includes: 1. the target state one-step prediction estimation of the next moment is obtained through a motion model basedon the target state optimal estimation of a k-1 moment; 2, the range information and angle information of the target relative to the radar are converted into the Cartesian coordinate position information of the target by the fixed-point sampling non-linear transformation method of random variable after the observation value of the target at the time k is obtained by the radar observation station.3, the two parts of the target state one-step prediction prior information and the radar observation inverse estimation likelihood function are combined through product fusion according to the probability likelihood product rule, and finally the posterior estimation of the target state at time k is obtained, and after the target state is stored, the moment is updated and the next iteration turn is entered. The invention has the characteristics of higher precision, better robustness and more concise algorithm structure, and has high practical value in radar, multi-sensor, maneuver and multi-target tracking.

Description

technical field [0001] The invention relates to a Bayesian filtering target tracking algorithm, which belongs to the field of target tracking. Background technique [0002] Target tracking has a wide range of applications in both military and civilian fields, such as air surveillance, satellite and spacecraft tracking, intelligent traffic and video surveillance, etc. The target tracking problem is essentially a state estimation problem, and its core is a filtering algorithm. [0003] According to the different space models of dynamic systems, filtering problems can be divided into linear filtering and nonlinear filtering. In the 1970s, the Kalman filter was successfully applied in the field of target tracking. As the most classic linear filtering algorithm in the field of target tracking, in the case of linear Gaussian, the filtering result of the Kalman filter is in the minimum variance, maximum likelihood are optimal under the same criteria. It can be proved by using th...

AI Technical Summary

Problems solved by technology

Facing the problem of nonlinear filtering, a large number of scholars and experts have proposed many effective nonlinear filtering algorithms in the past 30 years, the most famous ones are extended Kalman filter (EKF), unscented Kalman filter (Unscented Kalman filter). filter, UKF), particle filter (Particle filter, PF), but they all have some problems, such as the extended Kalman filter algorithm has low linearization accuracy and needs to calculate the complex Jacobian matrix; the general unscented Kalman filter algorithm exists Complicated calculations, filtering divergence and even distortion problems, particle filtering has problems of large calculation and weight degradation

Images