Technological advances in micro-electro-mechanical systemsMechanical systems (MEMS) have revolutionized healthcare monitoringMonitoring through the design of medical sensors which offer flexibility in the remote acquisition of patients’ physiological parametersPhysiological parameters. This paper therefore addresses the issue of combining several wireless biomedical sensorsBiomedical sensors to check on patients’ health. To achieve this, we analyze relevant studies on various scenarios of vital sign observation using multimodal data, focusing on criteria such as: the quality of the articles used, the multimodal data processingMultimodal data processing strategies, and the challenges involved. The period of articles selection for this study runs from January 1, 2018, to October 25, 2024. Based on PRISMA methodology, several databases, including IEEE, Web of Science, Arxiv, and PubMed, were used to identify 982 articles addressing our research question. We then selected articles according to inclusion and exclusion criteria. After filtering, 32 articles were selected. Our studies highlighted three approaches for remote healthMonitoring monitoringHealth monitoring using multimodal data: (1) building a specific device around a microcontroller integrating multiple sensors; (2) assembling wearable devices; and (3) using an open datasetDataset including several physiological parametersPhysiological parameters. Two data sources were identified: proprietary data and open datasets. Finally, form investigations, we found three concerns relating to the quality of the data used to perform validation, the relevance of signal fusion for predictionPrediction, and the lack of lightweightLightweight machine learningMachine learning models which can be deployed close to patients (Edge) for rapid feedback of health information.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Multimodal Sensing System for Health Monitoring: A Systematic Review

  • Awa Marah Nana,
  • Begonya Garcia-Zapirain

摘要

Technological advances in micro-electro-mechanical systemsMechanical systems (MEMS) have revolutionized healthcare monitoringMonitoring through the design of medical sensors which offer flexibility in the remote acquisition of patients’ physiological parametersPhysiological parameters. This paper therefore addresses the issue of combining several wireless biomedical sensorsBiomedical sensors to check on patients’ health. To achieve this, we analyze relevant studies on various scenarios of vital sign observation using multimodal data, focusing on criteria such as: the quality of the articles used, the multimodal data processingMultimodal data processing strategies, and the challenges involved. The period of articles selection for this study runs from January 1, 2018, to October 25, 2024. Based on PRISMA methodology, several databases, including IEEE, Web of Science, Arxiv, and PubMed, were used to identify 982 articles addressing our research question. We then selected articles according to inclusion and exclusion criteria. After filtering, 32 articles were selected. Our studies highlighted three approaches for remote healthMonitoring monitoringHealth monitoring using multimodal data: (1) building a specific device around a microcontroller integrating multiple sensors; (2) assembling wearable devices; and (3) using an open datasetDataset including several physiological parametersPhysiological parameters. Two data sources were identified: proprietary data and open datasets. Finally, form investigations, we found three concerns relating to the quality of the data used to perform validation, the relevance of signal fusion for predictionPrediction, and the lack of lightweightLightweight machine learningMachine learning models which can be deployed close to patients (Edge) for rapid feedback of health information.