Tracking object's type changes with fuzzy based fusion rule

TL;DR

This paper compares three data fusion rules—Dempster's, PCR5, and a fuzzy-based rule—analyzing their effectiveness in estimating target object types over time.

Contribution

It introduces a fuzzy-based fusion rule connecting traditional information fusion with fuzzy operators, and evaluates how different t-norms and t-conorms affect target type estimation.

Findings

Fuzzy-based fusion rule's performance varies with different t-norms and t-conorms.

PCR5 and Dempster's rule show distinct behaviors in type estimation.

Fuzzy operators can enhance or diminish the accuracy of target type tracking.

Abstract

In this paper the behavior of three combinational rules for temporal/sequential attribute data fusion for target type estimation are analyzed. The comparative analysis is based on: Dempster's fusion rule proposed in Dempster-Shafer Theory; Proportional Conflict Redistribution rule no. 5 (PCR5), proposed in Dezert-Smarandache Theory and one alternative class fusion rule, connecting the combination rules for information fusion with particular fuzzy operators, focusing on the t-norm based Conjunctive rule as an analog of the ordinary conjunctive rule and t-conorm based Disjunctive rule as an analog of the ordinary disjunctive rule. The way how different t-conorms and t-norms functions within TCN fusion rule influence over target type estimation performance is studied and estimated.