Target tracking method based on nonlinear model of Gauss-Newton filtering

Abstract

The invention discloses a target tracking method based on a nonlinear model of Gauss-Newton filtering. The method comprises the following steps that an optimal filtering model and the filtering memorylength are selected according to whether a target is maneuvered; model conditions are initialized by calculating a mixing probability and carrying out mixed estimation via the mixing probability; model conditions are filtered; the model probability is updated; and estimation is fused. Model condition filtering comprises that 1) a state is predicted primarily; 2) an residual error and a covariancematrix thereof are calculated, measured and predicted; and 3) filtering is updated by calculating a mixed covariance, determining a confirmation measurement condition if a confirmation port door, andsolving the volume of the confirmation port door. The provided target tracking method overcomes the disadvantages that a filter is easy to radiate and an algorithm is high in complexity with increaseof the clutter density when a present clutter tracking method is used to track a maneuvering target in the tracking clutter environment, and the tracking precision and stability are improved greatly.

Description

technical field [0001] The invention belongs to the technical field of target tracking, and in particular relates to a target tracking method under a nonlinear model based on Gauss-Newton filtering applied in a clutter environment. Background technique [0002] In the existing tracking methods for maneuvering targets in clutter environment, the filter is easy to diverge and the algorithm complexity is relatively high when the clutter density is high and the maneuverability is strong. The key to tracking the maneuvering target in the clutter environment lies in two points: one is the data association problem in the clutter environment, and the other is the mismatch between the state model and the actual motion model caused by the target maneuver. This method is based on the Gauss-Newton filter algorithm, making full use of the advantages of the Gauss-Newton filter algorithm for maneuvering target tracking, combined with the probabilistic strongest neighbor algorithm based on ...

AI Technical Summary

Problems solved by technology

[0003] The object of the present invention is to provide a target tracking method suitable for highly maneuverable targets in clutter environments, which solves the problems of poor stability, high track loss rate and algorithm complexity in the existing tracking methods for maneuver targets in clutter environments bigger problem

Images