Sistem Keamanan Brankas Senjata Api Komandan Berbasis Edge Computing dengan Autentikasi Biometrik Multimodal

Abstrak

Firearm inventory management in military environments demands robust security mechanisms to prevent unauthorized access and asset misuse. This study develops a security mechanism for a commander's pistol storage safe using a Multimodal Biometric Authentication approach integrated with Edge Computing architecture. Physical access is synchronously controlled via dual-biometric modality verification (inherence factors), namely fingerprint scanning and facial recognition. All biometric computations are executed on-device to eliminate external network dependencies, minimize latency, and mitigate cyber threats. Empirical testing on 20 test subjects (10 legitimate users and 10 impostors) across 100 experimental trials yielded a combined verification accuracy of 98.0% (comprising 49 true accepts and 49 true rejects). The system recorded an average processing time of 1.2 seconds from initial scanning to lock release, with an internal bus data-transfer latency of 15 ms. These findings demonstrate that integrating multimodal biometric authentication with edge computing delivers a self-contained, responsive, and precise physical security solution to safeguard the integrity and sovereignty of vital armaments in military command headquarters.