This project was created in collaboration between De Anza College & Infineon Technologies. The IntelliFan can automatically track a person and be controlled by gestures using the PSoC™ 6 AI Evaluation Kit and Raspberry Pi 4B.
Project Goal: Develop a smart fan that utilizes modern AI-driven human detection and gesture recognition to autonomously track and follow the user, providing a touch-free control experience.
- Autotracking of people in front of the fan
- Gestures such as swiping left / right / up / down, to control the fan
- Web-app so user can control fan with their phone / laptop
- 3D models to house the fan and its components, snap fit design
Person tracking is done through the Raspberry Pi Camera 3 and YOLO X Nano on NCNN for optimized performance, alongside Byte Tracking to generate consistent IDs across frames to distinguish individuals.
https://github.com/Qengineering/YoloX-Tracking-ncnn-RPi_64-bit
Person Tracking
IMG_0309.mov
Activate Person Tracking (Push)
IMG_0307.mov
Change Person (Left)
IMG_0313.mov
Change Person (Right)
IMG_0312.mov
Gesture detection is done through radar using the BGT60TR13C sensor on the PSOC 6 Ai Evaluation Kit, which is then fed into DeepCraft's radar gesture model.
https://github.com/Infineon/mtb-example-ml-deepcraft-deploy-ready-model
Turn On (Up)
IMG_0304.mov
Turn Off (Down)
IMG_0305.1.mov
Turn (Left)
IMG_0302.mov
Turn (Right)
IMG_0303.mov
The user can control the fan through their phone or laptop by connecting to the web server hosted over HTTP on the same WiFi.
Created by Warren and Zilu.
Web-app Demo
RPReplay_Final1742823615.mov
| Infineon PSoC™ 6 AI Evaluation Kit | $38.75 |
|---|---|
| Raspberry Pi 4 Model B (8GB RAM) | $39.99 |
| 32GB microSD Card | $9.99 |
| Raspberry Pi Camera Module 3 | $14.49 |
| PLA 3D-Printed Case and Housing | $6.99 |
| USB Type-C Cable | $4.99 |
| Servo Motor with Connecting Rod | $1.99 |