Overview
Barkbot was my final project for my Internet of Things course in spring 2024, where the brief was to apply our IoT skills to a machine of our choice. My machine had one very specific job: gather data on the barking. It listens to the room, shows how loud each bark is on a physical LED meter, and streams every reading to a live web page called the Barkbot Dashboard.
The Hardware
The brain is a Raspberry Pi Pico WH on a breadboard, programmed in MicroPython through Thonny. A SparkFun sound detector feeds its envelope output into one of the Pico's analog pins, and a ten segment LED bar graph, wired through a row of resistors with one pin per segment, works as a volume meter. The louder the sound, the more of the bar lights up.
From Bark to Dashboard
The Pico connects to WiFi and publishes readings over MQTT to a HiveMQ Cloud broker on two topics: sound_level, the raw number, and sound_rank, where that number lands on the meter. The dashboard is a very pink web page that graphs the live sound data with timestamps, so you can see exactly when the barking happened and how dramatic it got.
The Perpetrator: Peanut
Testing meant waiting for barks, and the source of those barks was Peanut, the small and extremely innocent looking dog in the gallery. The photo of her in my project folder is literally named bark_perpetrator, which tells you how the data turned out. She was a very good girl who happened to have a lot to say.
Peanut has since passed away. This project is a little bit of her, kept: a record of a machine I built to measure exactly how much she had on her mind. In memory of Peanut, the best and loudest collaborator I could have asked for.
The project started as a brainstorm page of sensor ideas, turned into a notebook sketch and a hand drawn wiring diagram, and grew through weekly progress notes, a class presentation, and a final reflection. I still love the grading philosophy of that course: you earned credit for documenting honest progress on your goals, not just for hitting every one of them.