DOG CARE SYSTEM

Dog care system with sound recognition, feeding mechanism and supporting web application.

About

A demo project that simulates a dog care system with the three functionalities above. Sound classifcation is done for the "bark", "growl" and "whine" sounds that can each be used to determine the dog's mental/physical state. Scheduling is done for deworming, feeding and grooming.

Project diagram

The servomotor can be attached to a dog feeder with rotating dispense. More info about the servo can be viewed on this repo.

Installation

1. Raspberry pi audio classifier setup

• Use edge impulse to train an audio classifier model.
• Create a mongodb atlas account and create a database.
• Follow the steps in the mongodb docs to obtain the connection string.
• Follow the steps in the servo control repo to setup the Raspberry Pi - Servomotor control.
• Clone the sound classifier repo on the pi and follow the steps in the readme. Open a terminal and install edge-impulse by running npm install edge-impulse-linux -g --unsafe-perm.
• Ensure SoX is installed.
• Download your model file from your Edge Impulse project/account with: edge-impulse-linux-runner --download modelfile.eim
• Connect a microphone and run node classify-audio.js modelfile.eim <microphone name>

2. Web app setup

• Clone this repo on your machine.
• Install required dependencies by running npm install in both the server and client folders.
• Obtain mongodb connection string and twilio api keys from mongodb atlas and Twilio
• In the server folder, create a .env file, add the following entries and replace:

ATLAS_URL = <your mongodb connection string>
TWILIO_ACCOUNT_SID = <your twilio sid>
TWILIO_AUTH_TOKEN = <your twilio auth token>

• Edit server.js and add your twilio details

function sendSms(message) {
  client.messages
      .create({
        body: message,
        from: 'YOUR TWILIO ACCOUNT NUMBER',
        to: 'TWILIO CONNECTED PHONE NUMBER'
      })
      .then(message => console.log(message.sid));
}

• Install nodemon globally by running npm install -g nodemon
• In the server folder run nodemon server.js
• In the client folder run npm start and navigate to https://localhost:3000

Pre-Requisites

• Edge impulse modelfile
• Node Js
• SoX