Multi-sensor quantitative fusion target tracking method based on variational Bayesian

Abstract

The invention relates to a multi-sensor quantitative fusion target tracking method based on a variational Bayesian method and a strong tracking information filtering method. According to the multi-sensor quantitative fusion target tracking method, a structure including a primary processor and a secondary processor are provided. In the primary processor, an enhanced measurement matrix H(k) and enhanced global information z (upsilon, k) are constructed; one-step prediction (k|k)-1) and corresponding covariance P(k|k-1) are calculated, global information predication z (k|k-1) is calculated, and a z (upsilon, k, 1), the (k|k)-1 and the P(k / k-1) are sent into the secondary processor. In the secondary processor, information noise variance is calculated, and (upsilon, k, 1) is sent to the primary processor, and in the primary processor, fusion estimation and corresponding covariance can be obtained through calculation, wherein please see the instruction for the formula of fusion estimation and corresponding covariance. Due to the fact that the variational Bayesian method and the self-adaptive strong tracking information filtering method are adopted in the multi-sensor quantitative fusion target tracking method, the high tracking capacity is achieved, unknown variance of noise can be estimated and measured, and the self-adaptive function can be achieved. Meanwhile, an attenuation coefficient can be estimated through an iterative method without calculating a jacobian matrix.

Description

technical field [0001] The invention belongs to the field of target tracking of a networked multi-sensor system, in particular to a multi-sensor information filtering quantization fusion method based on a joint variational Bayesian method and a strong tracking technology. Background technique [0002] In recent decades, due to the rapid development of modern information and Internet technology, more and more applications on complex wireless data networks have emerged. Therefore, a distributed sensor network composed of many distributed sensor nodes has been widely used in various fields, such as environmental monitoring, target tracking, intelligent fire protection, and health monitoring instruments and so on. Typically, data monitored by local sensors are quantified before they are transmitted to a processing center in order to meet data transmission and limited bandwidth requirements. For this reason, quantitative filtering and fusion have become research hotspots in sign...

AI Technical Summary

Problems solved by technology

Unfortunately, strong tracking cannot completely eliminate the influence caused by the uncertainty of the system model, and the quantization error variance cannot be estimated in real time

Images