Composite weighted fusion method and system with oscillation damping

Abstract

The invention relates to a composite weighted fusion method and system with oscillation damping, and the method comprises the steps of carrying out the weighting of a basic reliability distribution function of each group of evidences, and obtaining a first basic reliability distribution function of each group of evidences; obtaining a conflict discount coefficient between every two groups of evidences according to the first basic reliability distribution function of each group of evidences; performing composite weighted fusion on the first basic reliability distribution functions of every twogroups of evidences according to the plurality of conflict discount coefficients to obtain a basic reliability distribution function time sequence after composite weighted fusion; constaining random fluctuation of the basic reliability distribution function time sequence through oscillation damping constraint to obtain a corrected basic reliability distribution function time sequence, so that thecorrected basic reliability distribution function time sequence does not violate the common theory, the decision deviation and fluctuation generated when evidence priori knowledge is not fully utilized and conflicts exist among evidences are avoided, the performance is efficient and stable, theoretical derivation is complete, and the invention has good academic and engineering application prospects.

Description

technical field [0001] The invention relates to the technical field of multi-source information fusion, in particular to a compound weighted fusion method and system with oscillation damping. Background technique [0002] Evidence theory is a complete theory dealing with uncertainties, which can fuse different sensor data or classifier output results to obtain comprehensive decision-making under various assumptions; classical evidence theory has good information focus, but its There are three shortcomings: 1. It does not distinguish the confidence difference between evidences, resulting in inaccurate comprehensive decision-making; 2. When there are conflicts between evidences, there are usually unreasonable fusion results output; 3. Fluctuation constraints are not considered, resulting in The time series output of fusion probability is not stable; in summary, the classical evidence theory is prone to decision bias and fluctuations when the prior knowledge of the evidence is ...

AI Technical Summary

Problems solved by technology

[0002] Evidence theory is a complete theory dealing with uncertainties, which can fuse different sensor data or classifier output results to obtain comprehensive decision-making under various assumptions; classical evidence theory has good information focus, but its There are three shortcomings: 1. It does not distinguish the confidence difference between evidences, resulting in inaccurate comprehensive decision-making; 2. When there are conflicts between evidences, there are usually unreasonable fusion results output; 3. Fluctuation constraints are not considered, resulting in The time series output of the fusion probability is not stable; in summary, the classical evidence theory is prone to decision bias and fluctuations when the prior knowledge of the evidence is not fully utilized and when there is conflict between the evidence

Images