Biometric Fusion Strategies for Enhancing Access Control in Critical Infrastructures

Çakır H., Tanyeli O.

Turkish Journal of Mathematics and Computer Science, cilt.18, sa.2, ss.359-370, 2026 (Scopus, TRDizin)

Özet

Biometric systems enhance security using physiological and behavioral traits but face challenges like environmental factors and spoofing risks. Unimodal systems struggle with accuracy and reliability, prompting the adoption of biometric fusion. This strategy combines multiple modalities at different levels—sensor, feature, score, and decision—to improve robustness. This systematic review, following the PRISMA framework, analyzes 120 studies from 2000 to 2024 on fusion strategies for access control. It highlights performance advantages, key algorithms, and emerging trends, including deep learning, blockchain, and edge computing. The analysis demonstrates that multimodal systems significantly reduce error rates, such as achieving an EER of 0.85% compared to 2.5% in unimodal fingerprint systems. Furthermore, the review explores critical privacy-enhancing technologies and ethical concerns, such as data misuse and algorithmic bias, while proposing secure frameworks to mitigate these risks. The findings provide crucial insights for developing advanced, efficient, and secure biometric systems tailored for critical infrastructure protection. Future work should focus on explainable AI, standardization, and lightweight algorithms for broader adoption.