Case study · Product management
Making late night travel safer
RideSafe.AI is a driver-fatigue safety concept for ride-hailing: smarter shift limits, real-time drowsiness detection, and a driver fitness score passengers can actually see before they get in the car.
The problem
Night driving is the dangerous shift
6M
car crashes in the US every year
16,340
fatal car crashes in 2024
~50%
of those fatal crashes happened at night
Who we spoke to
Personas
Michael — driver
Late-night driver balancing multiple jobs. Struggles to recognise his own fatigue while feeling economic pressure to complete just one more ride.
Dev — passenger
Conversational rider who actively engages drivers on late trips to keep them alert, though he is often exhausted himself.
Jessica — passenger
Values quiet trips after a tiring schedule, but worries about driver alertness without wanting the responsibility of keeping them awake.
Solution · Part 1
Smarter shifts for safer driving
Drivers often work long shifts without breaks, increasing crash risk once fatigue sets in — especially at night.
- Step 1
Track driving hours
Backend logs ride times and shift durations.
- Step 2
Smart break triggers
After 3 hours, prompt an enforced 15-minute cooldown.
- Step 3
Ride lock + rest timer
The app disables ride acceptance during breaks.
Solution · Part 2
Real-time drowsiness detection
Drowsiness is detected in real time using machine learning with libraries such as Dlib and MediaPipe, then handled with a graduated response instead of a single blunt alarm.
Mild
In-app alert — a nudge to lower the window or try eye exercises.
Moderate
Ride acceptance temporarily blocked until the driver checks in.
Severe
AI call initiated to confirm the driver is fit to continue.
The payoff
A safety score you can trust
Shift logs and facial cues combine into a live driver fitness score. Passengers see a simple badge; behind it sits shift compliance, break history and drowsiness signals.
Data collection
Real-time monitoring of driver shifts and drowsiness. Shift logs plus facial cues make for safer decisions.
Real-time scoring
Score = breaks + drowsiness + shift compliance, turning raw data into a driver fitness score.
Live rating
Passengers see one badge — a well-rested driver earns trust at first ride.
How we'd measure it
Success metrics
Driver shift management
- Total shift duration
- Break compliance rate
- Overtime instances
Drowsiness detection
- Drowsiness events detected
- Yawning frequency
Driver scorecard
- Average driver fitness score
- Alert response rate
- Safe driving streaks
North star
- Building passenger and driver trust