Hello,
I would like to share my project, in which I use audio to detect LG Washing machine completion melody via ESP32, running DSP + Machine Learning. Integration with Home Assistant via MQTT. GitHub - mzakharo/lg_detector: Detect LG Appliance Completion with a Microphone, Machine Learning, and ESP32
ESP-DSP is used to generate a Spectrogram. esp-tflite-micro for Machine learning inference. Machine learning pipeline is generic, and the model can be re-trained to detect other sounds.
Interesting use of ML audio detection, rather than the traditional cutrent flow detection to monitor for cycle completion.
Why limit this to LG models? Nearly all washing machine models have some sort of alert they have finished their cycle.
You could make it smarter still and listen for the clothes dryer as well, as the two are usually co-located quite close to each other in the laundry.
If that doesn’t work, listening for when all the washing/drying noise ceases may be a viable option, as you have already purchased all the parts, and can pattern match to a cycle completion recording.
Your ‘laundry aid’…
If you could just automate taking the clothes out, checking for missing buttons, folding them neatly, and putting them in the correct location, and the wife, maid and butler could be redeployed for less mundane tasks.
This is fantastic work! Would it be easy to make it detect multiple sounds and report each one? For instance my top-load LG washer play a different tune when the washer is unbalanced during the spin cycle and needs attention. It would be very nice to have HA notify me in this case as well as when the load is done.
I wonder if it would it be possible to make this into a custom ESPHome component? MQTT is a great alternative, but it would be easier to configure the ESP via ESPHome which takes care of a lot of the infrastructure and enables additional capabilities.