Wesleyan Laundry Monitor
A deployed IoT system with 40+ active users on campus.
The Problem
Nobody wants to haul their laundry downstairs and find every machine taken.
Wesleyan students were walking to the laundry room with no way of knowing if a machine was free. I built a system that tells them before they leave. It runs on a physical device I designed, built, and installed myself. Within two weeks of launching, it had over 40 active users.
How It Works
01
Sense
A LIS3DHTR accelerometer picks up machine vibrations and relays the data to an ESP32 via I2C.
02
Classify
An on-device Random Forest classifier distinguishes running and idle states in real time.
03
Serve
Each node pushes its state to Firebase. The dashboard reads live from the database and every state change is logged, tracking usage per machine over time.
The Hardware
Microcontroller
ESP32, running the full classification pipeline on-device.
Sensors
LIS3DHTR accelerometer capturing vibration signatures unique to each machine state, communicating with the ESP32 over I2C.
Enclosure
Custom-designed housing with magnets on the back. Attaches directly to the machine with no modification. Completely non-invasive and removable in seconds.
Connectivity
Wi-Fi. Each node pushes state changes to Firebase Realtime Database, which the dashboard subscribes to directly.
ML Model
Random Forest classifier trained on labeled accelerometer data, runs entirely on-device without a server.
Battery Life
7+ months on a single charge, designed for unattended long-term deployment.
Real Usage
40+
Active Users
1
Building Live, More Coming
100%
On-Device Inference
7+
Months Battery Life
See It Live
weslaundry.com
Built With
Hardware
Software
This is a shipped product running on hardware I designed, built, and deployed solo. If you're looking for someone who can take a project from idea to production, I'd love to talk.
Get in Touch