Thesis Repository – 2025

This repository contains the implementation for the Social Robotics System for Senior Care using LLMs and Raspberry Pi.

---

🚀 Prerequisites

• Docker CLI (client version, not Desktop)

---

🛠️ Building the Project

Build and start the containers (first time may take up to 5 minutes):

docker compose up --build

If only code changes were made (no Dockerfile changes), you can just restart:

docker compose up

🤖 Installing Models (Ollama) Once the project is running, you need to install the desired model inside the Ollama container:

docker exec -it ollama bash
ollama pull mistral

To list installed models:

ollama list

or you can also browse available models at ollama.com/library

Client Setup

It is recommended to use a virtual environment for the client:

python3 -m venv client_env
source client_env/bin/activate

Install requirements:

pip install -r requirements.txt

Run the client:

python wspi_client.py

Model Configuration

STT (Speech-to-Text)

Change the model in stt_server.py:

return Model("models/vosk-model-small-en-us-0.15")

TTS (Text-to-Speech)

Change the model in tts_server.py:

tts = TTS(model_name="tts_models/en/vctk/vits").to(device)

Database Usage

Enter the Postgres container:

docker exec -it elderly_care_db bash

Connect to Postgres:

psql -U supauser -d elderly_care_db

List tables:

\dt

Configure IP on Client

Before starting up the client it is important to set correct server IP address:

hostname -I

The IP address needs to be then correctly set in .env file