Confidence transfer distributed volume Kalman filtering cooperative positioning method

Abstract

The invention discloses a distributed volume Kalman filtering cooperative positioning method based on confidence transfer. The method comprises the steps that the initial state of mobile agent nodes in a wireless sensor network meets Gaussian distribution, and the mobile agent nodes are cooperatively positioned and modeled into an edge posteriori distribution estimation problem of multivariable nodes in a time-varying factor graph; a Gaussian parameterization confidence transfer and reconstruction strategy and a distributed volume Kalman filtering method are proposed, the posteriori distribution of each variable node is calculated on the factor graph, and further a positioning result of each mobile agent node is obtained. According to the cooperative positioning method disclosed by the invention, only the Gaussian parameterization confidence coefficient needs to be transmitted between the adjacent agent nodes, the expansibility and robustness are good, the communication overhead and the calculation complexity are relatively low, and the overall network positioning precision and efficiency are improved.

Description

technical field [0001] The invention relates to the technical field of cooperative positioning of mobile nodes in a wireless sensor network, in particular to a distributed volumetric Kalman filter cooperative positioning method for confidence transfer. Background technique [0002] At present, under the background of the rapid development of unmanned navigation technology and mobile robot technology, the demand for location information is gradually increasing. The use of wireless sensor network cooperative positioning technology to realize mobile node positioning has received extensive attention and research from many scholars. The classic sum-product wireless network positioning algorithm uses a non-parametric confidence transfer strategy to realize cooperative positioning of agent nodes. In order to ensure better positioning accuracy, a large number of random sampling particles need to be applied, resulting in high computational complexity and also occupying a large number ...

AI Technical Summary

Problems solved by technology

The classic sum-product wireless network positioning algorithm uses a non-parametric confidence transfer strategy to realize cooperative positioning of agent nodes. In order to ensure better positioning accuracy, a large number of random sampling particles need to be applied, resulting in high computational complexity and also occupying a large number of sensors. resource

Since message passing between mobile nodes needs to send all particles representing position information, high communication overhead leads to high energy consumption, shortening the service life of sensors and batteries

The distributed extended Kalman filter algorithm is used to realize cooperative positioning. Due to the need to solve the Jacobian matrix, the high computational complexity limits its practical application, and the nonlinear function approximation strategy introduces model errors, resulting in limited positioning accuracy; The determination of parameters and weights in the Min Kalman filter algorithm will affect the final positioning result, and its covariance cannot always be guaranteed to be positive, resulting in unstable positioning results or even positioning failures

Images