Offline Local AI-Based Home Assistant for Early-Stage Dementia Self-Care Using Raspberry Pi 5
Specifications
- Code
- EA2610
- Department
- DKE
- Technology
- Raspberry Pi 5
- Theme
- Health & Assistive Care · AI & Computer Vision
- Student
- Adib Mujtaba bin Muhammad Mujada
- Supervisor
- Ir. Norhamimi bt Hamdan
Objectives
To develop an offline local AI-based home assistant using Raspberry Pi 5 to support daily self-care activities for individuals with early-stage dementia. To implement offline voice recognition and text-to-speech technologies, enabling users to interact with the system through spoken commands without internet connectivity. To design a user-friendly touchscreen interface that provides step-by-step visual and audio guidance for daily self-care activities.
Scope of project
The project focuses on assisting individuals with early-stage dementia through simplified instructions for daily activities. The system integrates Raspberry Pi 5, a touchscreen display, USB microphone, and speaker, using Vosk for offline speech recognition, Piper for speech synthesis, and Python Tkinter for the graphical interface. Users interact through the “Fiona” wake word and touchscreen controls to start activities, navigate between steps, repeat instructions, and return to the home screen without requiring internet connectivity.
Project description
“FIONA” is an offline local AI-based home assistant developed to support individuals with early-stage dementia in performing daily self-care activities. Powered by Raspberry Pi 5, the system integrates Vosk speech recognition, Piper text-to-speech, and a touchscreen interface to provide accessible voice and touch interaction. Users can activate the assistant by saying “Fiona” and receive step-by-step visual and audio instructions for activities such as preparing milk tea. The system operates locally without internet connectivity, supporting user privacy and accessibility. FIONA aims to encourage independent self-care while reducing reliance on continuous caregiver guidance.