ABSTRACT Multi‐sensor information fusion is a key technology for enhancing system perception and decision‐making capabilities. Random permutation set (RPS)‐based multi‐sensor information fusion algorithms can optimize multi‐source information integration while fully exploiting and utilizing the sequential features and structural information inherent in the data. However, when fusing highly conflicting RPSs using the permutation orthogonal sum, they often produce counterintuitive results. To address this limitation, this paper proposes a novel dual‐channel multi‐sensor information fusion algorithm that performs rule optimization and RPS modification simultaneously. The rule optimization channel employs an ordered permutation orthogonal sum, reconstructing the combination rules via ordered intersection operations and ordered conflict coefficients, which effectively exploits sequential constraints and precisely quantifies structural conflicts. The RPS modification channel introduces a novel divergence measure sensitive to variations in both element type and order, which effectively characterizes uncertainties arising from ambiguous information and facilitates optimal RPS weighting. The integrated design of these channels constitutes a comprehensive conflict management framework. To verify the algorithm's effectiveness, this paper applies it to three scenarios: threat assessment of computer viruses, iris classification, and fault mode and effects analysis of aircraft turbine rotor blades. Experimental results demonstrate that the proposed algorithm improves the average accuracy by in virus threat assessment, achieves an optimal ordered probability transformation value of 0.9909 in iris classification, and significantly enhances the reliability of turbine rotor blade risk ranking. These findings underscore the algorithm's substantial practical value.
Liu et al. (Wed,) studied this question.