Project Highlights:
- Used ARKit to track face and head position
- Built custom algorithms to analyze changes in the head position and eye movement
- Released the MVP within 2 months
iOS SOLUTION PREVENTING DROWSY DRIVING
Business Goals:
According to the National Sleep Foundation, each year around 100,000 traffic crashes in the USA are caused by drowsy driving, resulting in more than 6,400 deaths and 71,000 injuries. Due to our desire to prevent drivers from falling asleep while driving, we have developed the new MobiDev software product called WakeUp.
Product Description:
WakeUp is an iOS app that tracks the driver’s head position and eye movement in real-time. Once the app notices the signs of sleep on the face, it makes a sound alert to get the driver’s attention and keep them awake.
Applied Technologies:
Swift, ARKit, SceneKit, SwiftUI, TrueDepth camera, Apple Neural Engine, Agile management framework
Used ARKit to track face and head position
ARKit combines two sources of data receiving – a front-facing camera and a TrueDepth camera. Using them, ARKit detects facial feature points and eyelid closure level. Based on the collected parameters, the app constructs a 3D face model and can track the driver’s state with high accuracy.
Built custom algorithms to analyze changes in the head position and eye movement
For the collected tracking data, we have created a specific set of algorithms to analyze the head position and the state of eye position so that the app could determine whether the driver is about to fall asleep or not, and only after that produce a sound alert.
Released the MVP within 2 months
One of the biggest risks for an MVP launch is scope creep that makes the product team lose focus and endlessly add on features. Our killer feature was the combination of ARKit and custom algorithms, and we aimed to deliver it first. As a result, MVP took us 2 months. And now after the first version release, we will iteratively level up the app, in particular, following the requests we receive in users’ feedback.
ARKit
to track head position & eye movement
Custom algorithms
to increase the analysis accuracy
2 months
to deliver the MVP
