pico-factory

Demo project for RubyKaigi 2026 talk: PicoRuby for IoT: Connecting to the Cloud with MQTT

PicoRuby programs running on a Raspberry Pi Pico connect to AWS IoT Core over MQTT, send sensor data to the cloud, and display real-time detection counts on a browser dashboard. The dashboard also lets you send commands (start / stop / reboot) back to the device.

System Overview

[IoT Device (PicoRuby)]
     |
     | MQTT  pico-factory/device/{device_id}
     v
[AWS IoT Core]
     |
     | IoT Rule
     v
[DynamoDB: records]
     |
     | DynamoDB Streams
     v
[Lambda: stream_notifier]
     |
     | WebSocket push  {"count": N}
     v
[Browser Dashboard] <---> [API Gateway WebSocket]
     |
     | REST API  POST /publish, POST /clear
     v
[Lambda: api_handler] --> [AWS IoT Core] --> [IoT Device]

Repository Structure

pico-factory/
├── cloud/
│   └── pico-factory-backend/   # AWS SAM serverless backend
│       ├── template.yaml        # SAM template (all AWS resource definitions)
│       ├── samconfig.toml       # SAM CLI deployment config
│       ├── frontend/            # Browser dashboard (served via S3 + CloudFront)
│       ├── ws_handler/          # Lambda: WebSocket connection management
│       ├── stream_notifier/     # Lambda: DynamoDB Streams → WebSocket push
│       └── api_handler/         # Lambda: REST API (publish / clear)
└── device/
    ├── demo1.rb                 # PWM motor control only (no cloud)
    ├── demo2.rb                 # MQTT + motor control
    └── demo3.rb                 # MQTT + ToF distance sensor + motor control

Device Programs

File	Description
device/demo1.rb	Controls a DC motor via PWM. No cloud connectivity.
device/demo2.rb	Connects to AWS IoT Core over MQTT. Receives start / stop / reboot commands and drives the motor accordingly.
device/demo3.rb	Reads distance from a VL53L0X ToF sensor (I2C). Detects objects within 100 mm and publishes detection data to the cloud with a timestamp.

Tech Stack

Layer	Technology
Device	PicoRuby, MQTT, VL53L0X ToF sensor
Cloud	AWS IoT Core, DynamoDB, Lambda (Ruby 3.4), API Gateway, S3, CloudFront
Infrastructure	AWS SAM
Frontend	HTML/JS, Tailwind CSS, Chart.js, Luxon

Prerequisites

• AWS CLI with credentials configured
• SAM CLI
• Ruby 3.4
• Raspberry Pi Pico running PicoRuby

Quick Start

1. Deploy the backend

cd cloud/pico-factory-backend
sam build
sam deploy --guided   # first time only

After deployment, note the CloudFormation stack outputs:

Output key	Description
WebSocketUrl	WebSocket endpoint for the dashboard
RestApiEndpoint	REST API base URL for the dashboard
FrontendBucketName	S3 bucket to upload frontend files
FrontendUrl	CloudFront URL to open the dashboard

2. Upload the frontend

aws s3 sync ./frontend s3://{FrontendBucketName} --delete

3. Configure the dashboard

Open https://{FrontendUrl} and enter the WebSocketUrl and RestApiEndpoint in Settings.

4. Configure the device

Create a device certificate in AWS IoT Core, then update the following values in device/demo2.rb or device/demo3.rb:

ENDPOINT  = "xxxxxxxxxxxx-ats.iot.ap-northeast-1.amazonaws.com"
DEVICE_ID = "your-device-id"

Place the CA certificate, client certificate, and private key on the device filesystem and set their paths in the program.

Backend Details

See cloud/pico-factory-backend/README.md for full API reference, DynamoDB table design, local development guide, and teardown instructions.

Teardown

aws s3 rm s3://{FrontendBucketName} --recursive
sam delete --stack-name pico-factory-backend