Controlling Your Car With Gestures
PoC for a major German automotive company: A unique user interface with gesture recognition for control over the central console.
![](https://cdn.prod.website-files.com/64be79d12f4be686f39bcdee/66d19220e3f0468983c7b1b1_llm-big-data-success-story-motius-DE.webp)
![](https://cdn.prod.website-files.com/64bea1df65b74226fa13e905/64f576733070b0fe99f1170c_controlling-your-car-with-gestures-motius.webp)
Our Final Result
Overview
Challenge
The use case involved swiping your hand to receive or cancel calls with a SmartWatch as the device for gesture recognition. The information would be sent directly to the car.
Solution
We created a deep learning model to recognize some predefined hand gestures like swipe right/left, tap, etc. As a result, we converted that model to Apple’s own CoreML model that allowed us to run the neural network directly on the Apple Watch.
Technologies
- Keras
- CoreML
- Python
- TensorFlow
- WatchOS
Impact
This project was a technical Proof of Concept showcasing a unique user interface: gesture recognition while driving.Users were able to control the entertainment system of the car with simple hand gestures.