Reliability Analysis of AI-Enhanced Multimodal Assistive Systems for in-Home Elderly Care

View at publisher Summarise with VIRAT
Authors: Sinan Chen, Masahide Nakamura
Cited by: 0
References: 23

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This paper addresses the reliability challenges of AIassisted elderly care systems within smart home environments. By integrating multimodal inputs, including automatic speech recognition from instructional videos and optical character recognition from caregiving materials with context-aware processing, we develop a novel question answering system tailored for in-home care scenarios. A sliding context window mechanism prioritizes temporally relevant information, enhancing response accuracy. We propose three reliability metrics, Answer Accuracy Rate (AAR), Context Consistency Rate (CCR), and Modal Agreement Rate (MAR), to systematically assess system performance across four levels of care complexity, where MAR is reported separately as an auxiliary indicator of cross-modal consistency. Experimental results demonstrate optimal reliability at intermediate complexity levels, where procedures are well-structured and queries are clearly articulated. The findings contribute to the development of reliable AI-enhanced assistive systems, bridging advances in electrical engineering, automation technologies, and artificial intelligence for practical healthcare applications.

Fields of Study & Subjects

Technology Use by Older Adults Innovation in Digital Healthcare Systems Context-Aware Activity Recognition Systems

Publication details

Journal / source
International Conference on Electrical Automation and Artificial Intelligence (ICEAAI)
Pages
731–737
Publication date
16 January 2026
Language
EN