Design and Implementation of an Embedded System for Vehicle Anti-Theft Using Face Recognition and Live Location Tracking on Raspberry Pi

The rapid advancement of Information Technology, which has driven down vehicle prices, has led to a substantial increase in global vehicle sales over the past two decades. However, this rise in vehicle ownership has also been accompanied by a notable surge in car theft, creating a significant challenge for vehicle owners. As a result, a continuous battle exists between vehicle manufacturers and thieves. Unfortunately, traditional security methods often fall short when facing increasingly sophisticated vehicle theft techniques, which leads to the vehicle theft problem becoming a global issue. This paper proposes an Intelligent Vehicle Anti-Theft System (IVATS) as a solution for the vehicle theft problem. The proposed IVATS integrates GSM, GPS, Raspberry Pi, and a novel Challenge-Response Face Authentication model to improve the system security. The IVATS addresses limitations of existing solutions, such as reactive responses, unreliable user verification, and poor scalability, by integrating real-time tracking, remote mobile control, and multi-layered biometric authentication. The research contribution is threefold: (1) a high security framework leveraging IoT and embedded systems, (2) a challenge-response face authentication model incorporating emotion verification to deter spoofing, and (3) a user-friendly mobile app for remote monitoring. The IVATS employs a Raspberry Pi 3 and ATmega32 microcontroller to manage hardware modules (camera, GPS, GSM) and software components, including a Horizontal Ensemble Best N-Losses (HEBNL) model, which is a fine-tuned VGG16 model for emotion recognition. The challenge-response face authentication model achieves an average accuracy of 98.89% under optimal lighting but may degrade in low-light conditions and can authenticate users in 24 ms, which is suitable for the hardware devices that constitute the system. The integration of CRFA with FER outperforms traditional VATS by enabling dynamic, liveness-aware authentication based on real-time emotional responses, which substantially mitigates spoofing and replay attacks while preserving low-latency performance on embedded platforms. Moreover, IVATS offers a practical and efficient solution for modern vehicle security by effectively balancing strong robustness with user convenience.