Method of Improving the Data Fusion Accuracy of Multi-sensor System

Abstract

A method for improving the accuracy of data fusion of a multi-sensor system proposed by the present invention aims to provide a method that can effectively control the influence of local bad information on global estimation. The present invention is realized through the following technical scheme: in the local estimation component of the sensor system with relatively poor estimation accuracy or relatively large interference, a weighting factor λ is set, and matching parameters are set according to the λ calculation parameters α and β Fuzzy setter; each sensor independently collects the measurement points and passes through the local Kalman filter, and the obtained local state estimation 1-L input data is fused to the preprocessing module, and the data fusion preprocessing module estimates the error autocorrelation matrix according to the Kalman filter The definition of each local estimation error autocorrelation matrix is ​​extracted from the first row and the first column component, and the online solution is used to quantify the error ratio coefficient r, the error ratio coefficient change rate rc and the weighting factor λ that describe the relationship between each local estimation error. Adjust the value of λ; call λ to participate in the calculation during data fusion, and obtain the final global state estimation.

Description

technical field [0001] The invention relates to a multi-sensor system data fusion technology in the technical field of target tracking, and is a design method for improving tracking precision by adding an adaptive weighting factor of parameter fuzzy setting into a classical data fusion method. Background technique [0002] Target positioning and tracking is based on the principle of best estimation, using the calculation method of digital filtering, processing the measurement received by the sensor, and estimating the data processing process of the moving elements of the target. Measurements are noise-contaminated sensor observations about the state of a target, including other information such as slant range, azimuth, elevation, and time difference. Target motion elements generally refer to parameters such as target state and heading. The target state mainly refers to the motion components of the target (such as position, velocity, acceleration, etc.). Usually, target pos...

AI Technical Summary

Problems solved by technology

However, experiments have shown that P alone (i) (n) The adjustment strength is relatively limited. If the error of a certain sensor is particularly large, its estimation result will still participate in the weighted sum calculation, resulting in the loss of other better local estimation information, and the global estimation accuracy will inevitably be greatly reduced. pull down

Images