A Smart Self-Feeding System for Frail Elderly and Parkinsonian Patients

Abstract

This paper introduces a self-feeding machine that can help older people and people with Parkinson's disease eat independently, without having to rely on others for assistance. It consists of an index tray, a two-degree-of-freedom grip arm equipped with a spoon, and a mobile application for control, offering a cost-effective and accessible solution. The device uses machine learning to analyze nutritional intake, detect user mouth movements, and estimate calorie consumption. These capabilities enable personalized diet recommendations and real-time adjustments to tray positioning and feeding motions, thereby enhancing patient autonomy. By integrating efficient hardware design and nutritional analytics, CareFeeder represents a promising approach to improving both the independence and overall well-being of individuals with limited mobility.