AegisDrive: Event-Driven Driver Safety System
A comprehensive driver safety system that combines edge computing with cloud-native architecture to detect driver drowsiness in real-time. Built with C++ edge processing and an event-driven .NET 9 backend.
Project Demonstration
System Architecture
AegisDrive utilizes a hybrid edge-cloud architecture designed for high availability and low-latency safety alerts. The system balances local real-time AI inference on the edge with scalable data management and fleet-wide analytics in the cloud.
Edge Processing Intelligence
The Raspberry Pi 5 serves as the primary intelligence layer, executing concurrent pipelines for driver state analysis and road environment perception. This edge-based processing ensures deterministic, sub-100ms detection latency required for safety-critical alerts.
Software Logic Flow
Physical Hardware Unit
Custom 3D-printed enclosure optimized for concurrent AI inference and multi-camera heat dissipation.
The Challenge
The core engineering challenge was twofold: optimizing constrained edge hardware to run three concurrent AI models (Drowsiness, Head Pose, Road Detection) without lag, and ingesting a 'firehose' of high-frequency telemetry data into the cloud without overwhelming the database or creating dashboard latency.
The Solution
Implemented an event-driven microservices architecture that processes computer vision data at the edge (C++/OpenCV) and streams critical events to a .NET 9 cloud backend via AWS S3/SQS. This hybrid approach enables instant safety operator alerts while maintaining system-wide visibility and analytics capabilities.
Production Preview
Real-time Incident Report
Live dashboard showing driver drowsiness incidents with timestamp, severity, and location data for immediate safety operator response.
Live Operations Fleet Map
Geographical visualization tailored for both comprehensive fleet management and individual solo-driver monitoring. Upon destination entry, the system generates a dynamic blue route path with real-time ETA and distance tracking—similar to Google Maps—providing live updates on remaining minutes and kilometers until arrival.
Awards & Funding
Cairo AI Hackathon
Won and secured 50,000 EGP in investment funding at the Cairo AI Hackathon organized by Athar Accelerator. Competing against startups nationwide, AegisDrive was recognized for its innovative AI-powered road safety and driver monitoring solution.
UGRF Business Model Competition
Achieved 3rd place in the 20th International Junior Researcher Forum (UGRF) at Nile University. Secured 10,000 EGP in funding for presenting a scalable business model for event-driven fleet safety monitoring.