Hardware Track Finalist

In the middle of 2025, my team and I were looking around for competitions to boost our portfolio in preperation for our internships. I found an interesting looking one called "CodeNection 2025", which was a multi-month long hackathon with different tracks — one of which being hardware.

As the team genuinely loves tinkering around with hardware, we treated it as a sign to join. We soon found our problem statement where A/C units in campuses were often left running even when no one was using it. Following this, we designed an ecosystem to turn traditional IR-based A/C units into smart devices that could be controlled via an online dashboard.

In our original pitch, the ecosystem used a Raspberry Pi 5 with an INMP441 I2S microphone and a Raspberry Pi Camera Module 3 to detect human presence. The detections were done locally with a custom CNN for audio and YOLOv11n for image. As every A/C unit has its own IR code, we hooked up a KY-022 Infrared IR Receiver to capture these device-specific codes for later use.

In the end, the controller information is sent to an edge device on the A/C units, who is responsible for sending individual A/C details such as fan speed and temperature back to the Raspberry Pi 5, and finally triggering it with an IR blaster. All these information are also logged onto a web dashboard that allows for maximum customizations.

Our ecosystem estimated 47.8% decrease in energy usage for university campuses, largely due to its automated feature of ensuring an A/C stays off when no one is using it.