Skip to content

hyungminkimdev/HandDrive

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 

Repository files navigation

HandDrive

 WWDC23 Swift Student Challenge

Image

About the Project

This app, 'HandDrive', encourages driving for individuals with lower-body paralysis by using AR-based parking simulations to help them gain confidence and lead a more positive life.

Did you know that individuals with lower-body paralysis can also drive?

Many people with disabilities operate vehicles using assistive hand controls to manage acceleration and braking. However, despite the technical feasibility, psychological fears and anxieties often make them hesitant to take on the challenge.

HandDrive is an app designed to help these individuals experience parking simulations in an AR (Augmented Reality) environment. The app utilizes machine learning technology to recognize hand movements. The left hand, when clenched into a fist, controls steering motions, while the right hand’s gestures are used to manage acceleration and braking. This setup allows users to practice driving motions naturally, just as they would with actual assistive hand controls.

Through a three-stage parking simulation, users can practice driving, overcome perceived limitations, and build both understanding and confidence in their driving abilities.

Key Technologies & Implementations

Augmented Reality (AR) Development

  • ARKit: Built an immersive AR driving simulation that mirrors real-world scenarios, enhancing user engagement and realism.

Hand Gesture Recognition

  • Chirality Detection: Differentiated left and right hand interactions using hand.chirality == .left/.right.
  • Left Hand (Steering Control):
    • Implemented HandPoseClassification to simulate natural hand controller movements.
  • Right Hand (Acceleration & Braking Control):
    • Utilized HandActionClassification with a custom model, MyHandActionClassifier, for accurate motion control.

Machine Learning Optimization

  • CreateML: Trained distinct models for each hand gesture with over 95% accuracy through data augmentation techniques, ensuring reliability across diverse conditions.

Responsive Design & User Experience

  • Developed a grid-based layout optimized for various iPad devices, from Air to Pro, ensuring consistent performance and user comfort.
  • Crafted a seamless user journeyStart > Story > Walkthrough > Main > Ending—enhanced with modals and smooth animations for intuitive navigation.

Key Screenshots

Main View Story View Simulation View
Image Image Image

About

Driving Simulator for Lower Limb Disabilities

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors