Toward Harmonized Human–Machine Interaction: Assistive Communication for Elderly with Aphasia

Abstract

This research aims to develop a practical and user-friendly communication device for elderly individuals with higher brain dysfunction, particularly those with aphasia caused by stroke. Current gaze-tracking devices are often ineffective due to unstable eye movements in older users. To address this, the project integrates prior achievements in rehabilitation robotics and elderly monitoring to create functions for facial expression detection, classification, and adaptive learning. Unlike existing emotion-recognition systems, this approach seeks to infer communicative intent from facial features and translate it into speech and synchronized avatar actions, enhanced by adjusted voice frequency, speed, and gestures for clarity. The innovation lies in tailoring the system to individual disabilities, enabling urgent responses (e.g., breathing difficulties), and refining accuracy through machine learning. With Japan’s aging workforce—26% of 35 million seniors employed—communication support is increasingly vital to prevent workplace accidents and service decline. This study positions machines not as replacements but as partners harmonizing with human cognition and emotion. By combining assistive robotics expertise with monitoring avatars, the project aims for low-cost, socially implementable solutions applicable at home, in communities, and workplaces, ultimately contributing to safer and more inclusive elderly communication.

Keywords: Welfare Robotics, Rehabilitation Robotics

DOI: 10.54941/ahfe1007804