Theme:
Edge AI & Intelligent Systems
Do visual object recognition and send the result over LoRaWAN or NB-IoT
Bandwidth is scarce. Power is precious. The device must capture an image, run on-device object recognition, and transmit only the inference result over LoRaWAN or NB-IoT. Raw images may not be sent.
Every byte and every millijoule counts. Teams must optimize the AI model, compress the payload, and minimize transmission energy while maintaining classification accuracy.
How to win the challenge?
• Smallest payload sent
• Lowest total energy per inference + transmission
• Highest classification accuracy
Bonus: Real-time dashboard visualization.
Participants:
Deadline: August 31st