←Back to work

Case Study

SafeDrive AI

A real-time AI-powered speed monitoring app for iOS and Android that detects your current speed, fetches live speed limits from OpenStreetMap, and alerts drivers through voice, haptics, and visual cues when speeding.

Mobile App DevelopmentMobile App Developer2025
SafeDrive AI speed monitoring app banner

Project Highlights

  • Built a cross-platform React Native app with Expo and TypeScript that tracks GPS speed at 1-second intervals using expo-location, with a dual-calculation fallback combining sensor data and derived velocity.
  • Integrated the OpenStreetMap Overpass API to fetch live road speed limits within a 300m radius, parsing road class defaults, directional limits, school zone conditional tags, and time-based rules (e.g. school hours only).
  • Implemented custom haversine and bearing algorithms from scratch for road proximity detection and upcoming speed limit prediction with a 220m pre-warn distance — no external geo library used.
  • Designed a multi-modal alert system combining expo-speech voice TTS ('Reduce speed now'), expo-haptics, and visual state indicators (Safe / Warning / Alert) with configurable +5/+10/+15 km/h thresholds.
  • Built alert fatigue prevention logic requiring 4 seconds of sustained overspeed before triggering, plus a 25-second cooldown between alerts to avoid driver distraction.
  • Optimized performance using useRef-based tracking for non-render state, debounced API calls with 7-second and 70-metre movement thresholds, and a 35-second staleness cache for speed limit data.
  • Shipped cross-platform builds via EAS Build for iOS App Store and Android Play Store, with a demo mode that simulates drives without GPS for testing and development.