SAFERIDE AI

Abstract

Road accidents caused by driver fatigue, distraction, and unsafe driving behavior remain a significant global safety challenge. While advanced driver monitoring systems exist, they typically depend on specialized hardware and are often restricted to high-end vehicles, limiting their widespread adoption. Recent advances in mobile computing and artificial intelligence provide new opportunities to develop accessible and cost effective safety solutions using widely available devices. This paper presents SafeRide AI, a real-time driver monitoring system that transforms a smartphone or embedded camera into an intelligent in-car safety assistant. The proposed system employs computer vision techniques and facial landmark analysis to continuously evaluate driver behavior. By monitoring visual indicators such as eye closure patterns, head pose variations, and mouth movements, the system is capable of detecting critical conditions including drowsiness, distraction, yawning, and unsafe phone usage. When such behaviors are identified, the system generates immediate alerts to assist the driver in maintaining attention and reducing accident risk. The application is implemented using the Flutter framework for cross-platform deployment and integrates Google ML Kit for efficient face detection and landmark extraction. All processing is performed locally on the device, ensuring low detection latency, enhanced data privacy, and reliable offline functionality. The proposed approach offers a scalable, cost-effective, and easily deployable driver monitoring solution suitable for personal vehicles, commercial fleets, and smart transportation systems, demonstrating the potential of mobile-based AI to enhance real-world road safety.