Abstract

Smartphone anti-theft systems commonly provide device location, remote control, SIM-change monitoring, or camera-based evidence, but rarely combine these with biometric evidence. This study designed and evaluated a multimodal Android anti-theft system that combines Global Positioning System (GPS) tracking with photographic and fingerprint evidence acquisition and cloudbased remote monitoring. The system uses a three-layer architecture comprising an Android mobile layer, a Firebase cloud layer, and a React-based web layer. The Android application was developed in Kotlin using foreground services for location, camera and fingerprint operations. An external DigitalPersona U.are.U 4500 fingerprint scanner, connected via USB-OTG, acquired fingerprint samples and generated PNG images and Fingerprint Minutiae Data (FMD) templates in ANSI-378-2004 format. The front-facing camera and the GPS module captured and synchronised evidence with Firebase Realtime Database at five-second intervals. Firebase Storage held the photographic and fingerprint evidence, while a React/Leaflet web dashboard gave authorised users remote access to device location, captured evidence and tracking controls. Functional and performance testing confirmed successful operation of all major components under the defined test conditions. Observed response times were approximately 3 s for registration, 1.5 s for login, 3.5 s for scanner detection, 3 s for fingerprint capture response, 2 s for template extraction, 5–8 s for image upload, 5-s intervals for camera capture and GPS synchronisation, 10 s for dashboard loading, and 1 s for remote tracking control, with all services operating concurrently. These results demonstrate the technical feasibility of integrating location, visual and biometric evidence acquisition within a single Android anti-theft framework. However, the prototype remains dependent on external fingerprint hardware, Internet connectivity, device power, USB permissions and Android-device compatibility, and provides evidence representation rather than forensic-grade suspect identification. The system therefore offers a proof-of-concept for multimodal evidence collection rather than a complete forensic or law-enforcement solution.