ClassMind is an intelligent classroom management system built on RDK X5 that solves two real problems faced by every college, manual attendance and energy waste in empty classrooms.
Attendance System: A wide angle camera captures burst frames for 10-15 seconds. RDK X5 BPU runs YOLO face detection on each frame, tracks unique faces, splits classroom into zones, and selects the clearest image per student for recognition. Complete attendance for 50 students generated in under 20 seconds, zero teacher effort required.
Energy Management: Same camera monitors occupancy every 5 minutes. When zero people detected, RDK X5 signals ESP32 via ROS 2 to control relay switches for lights and fans automatically.
- Event: Robotics Dream Keeper Challenge by D-Robotics
- Track: Smart Life Robotics
- Timeline: June 1 – July 15, 2026
- Developer: Narendra Andhale
| Component | Purpose | Status |
|---|---|---|
| RDK X5 (4GB) | Main AI compute brain | Available |
| Wide angle MIPI camera | Face detection + occupancy | Pending |
| ESP32 | Relay control for lights/fans | Available |
| DHT22 sensor | Room temperature monitoring | Available |
| MQ2 gas sensor | Safety monitoring | Available |
| Relay module | Switch control | Available |
| Technology | Purpose |
|---|---|
| ROS 2 | System communication backbone |
| YOLO (Ultralytics) | Real time face detection on BPU |
| DeepFace | Face recognition and attendance |
| OpenCV | Camera capture and image processing |
| Python | Primary programming language |
| Ubuntu | Operating system on RDK X5 |
/camera_node
↓ raw frames
/burst_capture_node (10-15 sec, 20 frames)
↓ frame buffer
/zone_splitter_node (6 zones)
↓ zone images
/detection_node (YOLO on BPU)
↓ detected faces per zone
/recognition_node (DeepFace)
↓ student IDs
/attendance_node
↓ attendance report
/energy_node (occupancy monitoring)
↓ occupancy status
/esp32_bridge_node
↓ relay commands
Lights + Fan Control
classmind-rdkx5/
├── README.md
├── src/ ← Python source code
├── docs/ ← Documentation
│ ├── PROPOSAL.md
│ ├── ROADMAP.md
│ ├── STAGE1.md
│ └── DISCORD_POST.md
├── hardware/ ← BOM and wiring
│ └── BOM.md
├── assets/ ← Screenshots and evidence
└── launch/ ← ROS 2 launch files
- Application form submitted
- Registration confirmed by D-Robotics
- GitHub repository created
- Discord joined — username: naren
- Project concept defined — ClassMind
- System architecture designed
- OpenCV installed and working on laptop
- Face detection running on laptop (Haar Cascade)
- Live webcam detection working
- Group photo multi-face detection working
- RDK Studio registered
- YOLO installed and running — 86% confidence
- DeepFace face recognition working
- RDK X5 board received
- Camera connected to RDK X5
- YOLO running on BPU
- ESP32 relay connected
- Full ROS 2 pipeline working
- Deployed in IIIT Nagpur lab
- Stage 1 submitted
- Stage 2 submitted
- Stage 3 submitted
- Registered for Robotics Dream Keeper Challenge
- Joined Discord community
- Created GitHub repository
- Learned complete OpenCV fundamentals
- Built working face detection using Haar Cascade
- Live webcam detection working
- Multi-face detection on group photo working
- YOLO person detection working ,86% confidence
- RDK X5 shipping confirmed from D-robotics, arrives ~June 18
- Successfully flashed RDK OS 3.5.0 Desktop on RDK X5
- Connected the board to Wi-Fi and verified internet connectivity
- Enabled SSH and remotely accessed the board from laptop
- Explored the RDK Model Zoo and runtime examples
- Learned how the BPU inference pipeline works on RDK X5
- Used DroidCam as a temporary vision sensor during development
- Successfully ran YOLO11 object detection on the RDK X5 BPU
- Performed custom image inference and verified detection results
- Submitted Stage 1 Pull Request to the Robotics Dream Keeper Challenge
- Challenge: Robotics Dream Keeper Challenge by D-Robotics
- Official Repo: https://github.com/D-Robotics/Robotics-Dream-Keeper-Challenge
- Discord: D-Robotics Community (username: naren)
- Developer: Narendra Andhale — IIIT Nagpur
B.Tech ECE IoT student at IIIT Nagpur (2024-2028). Co-Head of IoTics Club Robotics Wing. Experience with ROS 2, Raspberry Pi, ESP32, Arduino, Python, and embedded systems. Previously built a quadcopter drone, weather station, and ROS 2 target tracking system. Interned at India Space Lab.
ClassMind — Making classrooms intelligent, one frame at a time.