Tracking filtering method based on data fusion

Abstract

The invention provides a tracking filtering method based on data fusion. The method comprises the following steps: firstly, calculating each Sagna sampling point and a corresponding weight value; propagating each Sigma sampling point through a nonlinear function; obtaining a mean value estimation and a covariance estimation of an observation vector, decomposing a nonlinear system into nonlinear filters with a plurality of subsystems, carrying out state estimation based on local filtering of an insensitive Kalman filtering method, and obtaining Kalman filter state estimation based on a data fusion algorithm after calculating the weight of each local filter. According to the obtained filtering estimation value, the estimation precision of the filtering algorithm can be improved, and the reliability of a filtering system is also improved.

Description

technical field [0001] The invention belongs to the field of directional positioning and relates to a tracking filtering method. Background technique [0002] The Kalman filter method is a very important method in passive tracking, and has been widely used in military and civilian fields. The traditional Kalman filtering method is prone to error accumulation during the filtering process, and when the initial state value and initial covariance errors are large, it is easy to cause filtering divergence. [0003] The insensitive Kalman filter method does not need to linearize the nonlinear system, and can be easily applied to the state estimation of the nonlinear system. However, the insensitive Kalman filtering method has some disadvantages, such as being easily affected by the initial state error and observability, because the insensitive Kalman filtering method uses a Gaussian distribution to approximate the posterior probability distribution of the system state, In the ca...

AI Technical Summary

Problems solved by technology

However, the insensitive Kalman filtering method has some disadvantages, such as being easily affected by the initial state error and observability, because the insensitive Kalman filtering method uses a Gaussian distribution to approximate the posterior probability distribution of the system state, In the case where the posterior probability distribution of the system state is non-Gaussian, the filtering result will have a huge error

Images