Open AccessProceeding Paper

CareTaker.ai—A Smart Health-Monitoring and Caretaker-Assistant System for Elder Healthcare^†

Ankur Gupta

Sahil Sawhney

^2,* and

Suhaib Ahmed

Department of Computer Science and Engineering, Model Institute of Engineering and Technology, Jammu 181122, Jammu and Kashmir, India

School of Management, Model Institute of Engineering and Technology, Jammu 181122, Jammu and Kashmir, India

Centre for Research, Innovation and Entrepreneurship, Model Institute of Engineering and Technology, Jammu 181122, Jammu and Kashmir, India

Author to whom correspondence should be addressed.

^†

Presented at the 1st International Conference on AI Sensors & The 10th International Symposium on Sensor Science, Singapore, 1–4 August 2024.

Eng. Proc. 2024, 78(1), 7; https://doi.org/10.3390/engproc2024078007

Published: 8 January 2025

(This article belongs to the Proceedings of The 1st International Conference on AI Sensors & the 10th International Symposium on Sensor Science)

Download

Browse Figures

Versions Notes

Abstract

There are several systems for patient care, including elderly healthcare, which rely on sensor data acquisition and analysis. These sensors are typical vital-monitoring sensors and are coupled with Artificial Intelligence (AI) models to quickly analyze emergency situations or even predict them. These systems are deployed in hospitals and require expensive monitoring and analysis equipment. Eldercare specifically encompasses monitoring, smart analysis, and even the emotional aspects of care. Existing systems do not provide a portable, easy-to-use system for at-home eldercare. Further, existing systems do not address advanced analysis capabilities around mood/sentiment/mental state/mental disorder analysis or the analysis of issues around sleep disorders, apnea, etc., based on sound capture and analysis. Also, existing systems disregard the emotional needs of elderly patients, which are a critical aspect of patient wellbeing. A low-cost and effective solution is therefore required for extended use in eldercare. In this paper, the CareTaker.ai system is proposed to address the shortcomings of the existing systems and build a comprehensive caretaker assistant using sensors, audio, video, and AI. It consists of smart bed sheets, pillow covers with embedded sensors, and a processing unit with GPUs, conversational AI, and generative AI capabilities, with associated functional modules. Compared to existing systems, the proposed system has advanced monitoring and analysis capabilities with potential for low-cost mass manufacturing and a widespread commercial application.

Keywords:

eldercare; elderly healthcare; healthcare monitoring; AI-based eldercare; eldercare smart assistant; smart healthcare system

1. Introduction

As per the United Nations Population Fund (UNFPA) report, the number of elder persons aged over 60 years is expected to reach two billion by 2050 [1]. Many of these elderly persons experience some sort of disability which confines them to beds for longer intervals of time. Elderly patients with chronic conditions such as diabetes, hypertension, or cardiovascular disease become more vulnerable to complications such as pressure ulcers, respiratory issues, and deep vein thrombosis which may arise from extended immobility. This poses a challenge for the healthcare systems in many countries in the world.

Traditional healthcare solutions [1,2,3,4,5,6,7] often rely on caregivers or wearable devices for monitoring and tracking the vital signs of bedridden and eldercare patients. However, these systems and methods have some limitations. Wearable devices can cause discomfort over time, particularly for patients with sensitive skin or cognitive impairments. Furthermore, manual checks and interventions fail to provide continuous and real-time monitoring. This lack of continuous and non-intrusive monitoring leads to delayed heath condition detection, increased healthcare costs, and a poor emotional experience for the patients.

Hence, there is a need for an automated, real-time health monitoring solution that addresses the challenges faced by physically disabled and bedridden individuals. To address this, a system and method for monitoring the health of users with physical disabilities and those requiring eldercare is presented in this paper. The system integrates sensors embedded in everyday items such as bed sheets and pillow covers. This offers a non-intrusive way to continuously monitor various physiological parameters including temperature, pressure, moisture, pulse rate, blood pressure, oxygen saturation, and movement patterns. In addition, the system uses generative AI which allows for more personalized care. It adapts to the personalized health profiles and needs of individual patients. It can also track long-term trends in health data and provide predictive insights that traditional monitoring tools may not be able to provide.

This research thus contributes to ongoing work on healthcare automation and AI in medical monitoring. It provides a new framework for addressing the specific needs of bedridden and elderly individuals, making it a valuable tool for caregivers, healthcare professionals, and patients. The rest of the paper is organized as follows: Section 2 presents a review of various healthcare monitoring systems and their limitations. The proposed system architecture and methodology is presented in Section 3, followed by discussions of future work in Section 4 and a conclusion in Section 5.

2. Literature Review

Numerous papers have been presented in the literature focused on healthcare monitoring systems [1,2,3,4,5,6,7]. Table 1 summarizes some of these health monitoring systems. Despite the advancements in health-monitoring technologies, these systems still face several limitations when applied to physically disabled and elderly patients. Some of these limitations include the following:

Lack of Comprehensive Monitoring: Existing solutions often focus on specific health metrics (for example cardiovascular, respiratory) and do not provide a detailed view of the patient’s health. This limits their ability to predict multi-factored health issues that bedridden or disabled patients face.
Limited Predictive Capabilities: While AI-based systems exist in the literature, many of them do not provide real-time predictive analysis. There is also a lack of AI integration for non-invasive sensor networks specially for the disabled individuals.
Invasive Devices: The majority of the health-monitoring devices require regular manual operation and are uncomfortable for long-term use. This makes them unsuitable for physically disabled patients who need continuous, hands-free, and automated monitoring.

The proposed system bridges these gaps by offering a comprehensive, AI-powered solution. Furthermore, the existing systems do not address advanced analysis capabilities around mood, sentiment, mental state, or mental disorder analysis, or the analysis of issues around sleep disorders and apnea, based on sound capture and analysis. Also, existing systems disregard the emotional needs of elderly patients, which are a critical aspect of patient wellbeing. A low-cost and effective solution is therefore required for extended use in eldercare.

3. Methodology

The steps used to design and develop an eldercare health monitoring system are described in this section. System design, hardware integration, AI model creation, data processing, and system validation are the stages of methodology that make up this caretaker system.

3.1. System Architecture

The proposed system for monitoring the health of elders and patients provides automated real-time health tracking and predictive analysis. The architecture of the proposed system is divided into three main categories which are hardware components (sensors, input/output devices), an AI-based processing engine, and a communication network. These components, shown in Figure 1, work together to continuously track the health of users and report it in real-time.

Embedded Sensor Network: The network consists of several physiological sensors such as a temperature sensor, moisture sensor, pressure sensor, oxygen saturation sensor, movement sensor, and pulse and blood pressure sensor. These sensors are embedded in the sheets and pillows of the bedding to monitor the vital signs of the patient and detect their movements.
AI-Driven Central Processing Unit (CPU): The AI engine is responsible for processing the data collected by the sensors and analyzing real-time physiological changes. Based on these data, this AI engine predicts potential health issues, if any.
User and Caregiver Interaction Interface: Input and output devices, such as speakers, microphones, cameras, and display devices, enable the system to interact with the user through voice recognition and generate reports and alerts for caregivers or healthcare providers.

3.2. Hardware Integration

The hardware components of the system are integrated for non-invasive and effective health monitoring. Multiple sensors are embedded into the patient’s bedding, thereby eliminating the need for wearable devices or continuous physical user intervention. Temperature and pressure sensors can be embedded in key areas of the bedding such as beneath the back, legs, and head, where prolonged immobility often results in pressure sores. Moisture sensors can be positioned in areas prone to sweating, whereas the blood pressure and oxygen saturation sensors can either be embedded in the bedding or included as part of unobtrusive wearable elements.

The system will be configured to collect data at predefined intervals, for instance every five minutes. This frequency can be adjusted as per the requirements. These sensors operate on a real-time data transmission model to continuously send physiological data to the CPU. The communication network facilitates this real-time data transmission between the sensors, the AI engine, and the caregivers or healthcare professionals.

The system uses the following technologies to ensure seamless operation.

Wireless Data Transmission: Data collected by the sensors is wirelessly transmitted to the processing engine. Depending on the user’s environment, the system can use Wi-Fi or Bluetooth for this.
Cloud Integration: The system is integrated with cloud platforms that allow the remote access of health data by the caregiver or healthcare providers.
Mobile App Connectivity: Alerts, reports, and health summaries can be sent to family members and caregivers via a mobile application for the remote monitoring of the patient.

Figure 2 shows the individual components of the proposed eldercare system. The system includes interactive features framed to enhance the user’s mental and emotional wellbeing. The system can play music, engage in conversations using pre-programmed dialogs, or remind the user about medication schedules.

3.3. AI Engine

The AI engine, consisting of graphical processing units (GPUs), is at the core of the CareTaker.ai system’s decision-making process. It ensures that collected data are not only monitored but also analyzed in real-time. This enables us to detect patterns and predict potential health issues before they become severe. The following operations will be performed by this AI engine:

Data Analysis: After data acquisition, a filtering process will be applied to remove irrelevant data. For example, random movements can be excluded from the analysis. The AI engine processes this filtered data to analyze physiological parameters such as blood pressure fluctuations, body temperature trends, and movement patterns by using feature extraction. Machine learning algorithms are used to identify any abnormalities in health metrics.
Pattern Recognition: The system is equipped with pattern recognition capabilities based on the historical data present in the database. For example, the AI can detect patterns of abnormal pressure distribution in the patient that indicate the formation of pressure ulcers, or irregular breathing patterns that could signal respiratory distress.
Predictive Alerts: The AI engine sends predictive alerts to caregivers and healthcare professionals which allows for early interventions by the caregivers or healthcare personals. These alerts will be delivered via a mobile application or caregiver dashboard, ensuring timely notifications and reducing the risk of delayed interventions. The alert also contains detailed information about the condition, for example, a high risk of a pressure ulcer on the lower back due to prolonged immobility for 3 h and recommends an action like changing the user’s position. Figure 3 (adopted from [15]) illustrates one such mechanism for movement alert process.
Adaptive Learning: To ensure the caretaker system becomes more accurate in its health predictions for the patient, the AI model can continuously learn from new data and refine its predictions over time.
Interactive Experience: To provide an interactive experience for the patient, the AI engine may generate one or more games and sentences for one or more conversations with the patient. Through voice commands, the AI engine can also play music as per requests via the one or more output devices. The AI engine can also connect the primary user with caretakers via voice calls.

The system also generates daily, weekly, and monthly reports summarizing the user’s health status. These reports include data on vital signs, movement patterns, and any detected anomalies. Healthcare providers can access these reports remotely through a secure cloud platform, ensuring continuous care even when they are not physically present.

4. Results

Night-time EEG recordings from the dataset [16,17] were taken for this study. The dataset comprised EEG recordings which were obtained at a sampling frequency of 512 Hz from patients exhibiting Periodic Limb Movement Syndrome (PLMS). The participants in the dataset were between the ages of 14 and 82, with an average age of 45. Of the study’s total respondents, 42% (44) were female and 58% (62) were male. A 60:40 training and testing ratio of the dataset was used in our analysis, and the sleep pattern was divided into five cycles (sleep stages) shown in Figure 4.

To ensure the accuracy and reliability of the proposed system, extensive testing and validation are needed in both simulated environments and real-world case studies. Initially, a simulation environment was established as a part of the preliminary work. Simulations were performed on the following computational setup: Dell Precision 3630 Tower Workstation with Xeon E2146G 3.5 GHz Processor, 32 GB DDR4 RAM, 512 GB NVMe SSD and NVIDIA Quadro M4000 8 GB graphic card (procured from Dell, Mumbai, India). Support Vector Machine (SVM) [18], Ensemble Bagged Tree (EBT) [19], and k-Nearest neighbors (KNN) [20] learning models were used in the AI engine for the classification of sleep disorder periodic leg movements during sleep (PLMS). The comparative analysis of these models is presented in Figure 5. It is observed that SVM achieves the highest accuracy of 94% in the classification of PLMS disorder.

The proposed CareTaker.ai system presents several advantages over existing solutions, such as the following:

Automated Monitoring: The continuous real-time data collection reduces the need for human intervention.
Predictive Analysis: The AI engine provides early warnings for potential health risks, if any, thereby improving patient outcomes.
User Interaction: The system offers interactive entertainment features using generative AI for patient engagement, which improves patient mental health.

Table 2 gives a comparison of sensor-based monitoring systems. The proposed caretaker system is in its initial development stages and in future the following work will be included. (1) Enhancements to AI Capabilities: more AI models can be explored to achieve higher prediction accuracy, especially in the detection of medical conditions which are not common. (2) Expansion to Rural Remote Healthcare: the system could be expanded to include remote healthcare capabilities and aid consultation in rural areas. (3) Integration with Wearable Devices: further integration with wearable health devices such as smartwatches or glucose monitors could provide more comprehensive health-monitoring solutions.

5. Conclusions

In this paper, the CareTaker.ai system is proposed to address the shortcomings of existing healthcare systems. It is a comprehensive caretaker assistant, using sensors, audio, video, and AI. It consists of smart bed sheets, pillow covers with embedded sensors, and a processing unit with GPUs, conversational AI, and generative AI capabilities, with associated functional modules for monitoring the health of users with physical disabilities. By leveraging sensor technology and AI-driven analysis, the system automates the detection of health risks, enhancing both patient care and caregiver support. The preliminary results demonstrate the effectiveness of the adopted SVM model in predicting risks with 94% accuracy. This system shows a strong potential for future development and widespread adoption in healthcare settings.

Author Contributions

A.G., S.S. and S.A. contributed equally to conceptualization, methodology, writing—original draft preparation, and writing—review and editing. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest.

References

Kuo, M.-H.; Wang, S.-L.; Chen, W.-T. Using information and mobile technology improved elderly home care services. Health Policy Technol. 2016, 5, 131–142. [Google Scholar] [CrossRef]
Ianculescu, M.; Alexandru, A.; Paraschiv, E.-A. The Potential of the Remote Monitoring Digital Solutions to Sustain the Mental and Emotional Health of the Elderly during and Post COVID-19 Crisis in Romania. Healthcare 2023, 11, 608. [Google Scholar] [CrossRef] [PubMed]
Raja, G.B.; Chakraborty, C. Internet of things based effective wearable healthcare monitoring system for remote areas. In Implementation of Smart Healthcare Systems Using AI, IoT, and Blockchain; Elsevier: Amsterdam, The Netherlands, 2023; pp. 193–218. [Google Scholar]
Lin, C.-C.; Chiu, M.-J.; Hsiao, C.-C.; Lee, R.-G.; Tsai, Y.-S. Wireless health care service system for elderly with dementia. IEEE Trans. Inf. Technol. Biomed. 2006, 10, 696–704. [Google Scholar] [CrossRef] [PubMed]
Yadav, R.; Pradeepa, P.; Srinivasan, S.; Rajora, C.S.; Rajalakshmi, R. A novel healthcare framework for ambient assisted living using the internet of medical things (IOMT) and deep neural network. Meas. Sens. 2024, 33, 101111. [Google Scholar] [CrossRef]
Yazici, A.; Zhumabekova, D.; Nurakhmetova, A.; Yergaliyev, Z.; Yatbaz, H.Y.; Makisheva, Z.; Lewis, M.; Ever, E. A smart e-health framework for monitoring the health of the elderly and disabled. Internet Things 2023, 24, 100971. [Google Scholar] [CrossRef]
Kumar, D.; Sood, S.K.; Rawat, K.S. Empowering elderly care with intelligent IoT-Driven smart toilets for home-based infectious health monitoring. Artif. Intell. Med. 2023, 144, 102666. [Google Scholar] [CrossRef] [PubMed]
Debache, I.; Jeantet, L.; Chevallier, D.; Bergouignan, A.; Sueur, C. A lean and performant hierarchical model for human activity recognition using body-mounted sensors. Sensors 2020, 20, 3090. [Google Scholar] [CrossRef]
Hasan, M.K.; Shahjalal, M.; Chowdhury, M.Z.; Jang, Y.M. Real-time healthcare data transmission for remote patient monitoring in patch-based hybrid OCC/BLE networks. Sensors 2019, 19, 1208. [Google Scholar] [CrossRef] [PubMed]
Gibson, R.M.; Amira, A.; Ramzan, N.; Casaseca-de-la-Higuera, P.; Pervez, Z. Matching pursuit-based compressive sensing in a wearable biomedical accelerometer fall diagnosis device. Biomed. Signal Process. Control 2017, 33, 96–108. [Google Scholar] [CrossRef]
Hassan, M.M.; Gumaei, A.; Aloi, G.; Fortino, G.; Zhou, M. A smartphone-enabled fall detection framework for elderly people in connected home healthcare. IEEE Netw. 2019, 33, 58–63. [Google Scholar]
Medrano, C.; Plaza, I.; Igual, R.; Sánchez, Á.; Castro, M. The effect of personalization on smartphone-based fall detectors. Sensors 2016, 16, 117. [Google Scholar] [CrossRef] [PubMed]
Kakria, P.; Tripathi, N.; Kitipawang, P. A real-time health monitoring system for remote cardiac patients using smartphone and wearable sensors. Int. J. Telemed. Appl. 2015, 2015, 373474. [Google Scholar] [CrossRef] [PubMed]
Chen, S.-T.; Lin, S.-S.; Lan, C.-W.; Hsu, H.-Y. Design and development of a wearable device for heat stroke detection. Sensors 2017, 18, 17. [Google Scholar] [CrossRef] [PubMed]
Tapwal, R.; Misra, S.; Deb, P.K. I-sheet: A low-cost bedsheet sensor for remote diagnosis of isolated individuals. IEEE Sens. J. 2022, 23, 906–913. [Google Scholar] [CrossRef] [PubMed]
Goldberger, A.L.; Amaral, L.A.; Glass, L.; Hausdorff, J.M.; Ivanov, P.C.; Mark, R.G.; Mietus, J.E.; Moody, G.B.; Peng, C.-K.; Stanley, H.E. PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. Circulation 2000, 101, e215–e220. [Google Scholar] [CrossRef]
Terzano, M.G.; Parrino, L.; Smerieri, A.; Chervin, R.; Chokroverty, S.; Guilleminault, C.; Hirshkowitz, M.; Mahowald, M.; Moldofsky, H.; Rosa, A. Atlas, rules, and recording techniques for the scoring of cyclic alternating pattern (CAP) in human sleep. Sleep Med. 2002, 3, 187–199. [Google Scholar] [CrossRef] [PubMed]
Huang, S.; Cai, N.; Pacheco, P.P.; Narrandes, S.; Wang, Y.; Xu, W. Applications of support vector machine (SVM) learning in cancer genomics. Cancer Genom. Proteom. 2018, 15, 41–51. [Google Scholar]
Wang, Z.; Wang, Y.; Srinivasan, R.S. A novel ensemble learning approach to support building energy use prediction. Energy Build. 2018, 159, 109–122. [Google Scholar] [CrossRef]
Zhang, S.; Li, X.; Zong, M.; Zhu, X.; Cheng, D. Learning k for knn classification. ACM Trans. Intell. Syst. Technol. (TIST) 2017, 8, 43. [Google Scholar] [CrossRef]
Choi, Y.; Jeon, Y.-M.; Wang, L.; Kim, K. A biological signal-based stress monitoring framework for children using wearable devices. Sensors 2017, 17, 1936. [Google Scholar] [CrossRef]
Hossain, M.S.; Muhammad, G. Cloud-assisted industrial internet of things (iiot)–enabled framework for health monitoring. Comput. Netw. 2016, 101, 192–202. [Google Scholar] [CrossRef]

Figure 1. Overview of caretaker system.

Figure 2. Block diagram representing the components and modules of the proposed eldercare system.

Figure 3. Illustration of movement alert in the caretaker system.

Figure 4. Different sleep stages detected from EEG signals.

Figure 5. Comparison of different AI models.

Table 1. Literature review summary.

Ref.	Description	Methodology	Sensors Used	Features
[8]	Human activity recognition (HAR)	Machine learning	Accelerometer and gyroscope	Low-complexity model with high accuracy
[9]	Wearable ECG monitoring	Transmission of data through hybrid infrastructure	ECG and cameras	Efficient and reliable data transmission
[10]	Fall diagnosis	Data collection, processing, and classification	Accelerometer	Low battery consumption, accurate but no security
[11]	Fall detection	Real time detection with smartphone alerts	Accelerometer	Good accuracy
[12]	Personalized fall detection	Machine learning	Smart phone with accelerometer	Good accuracy with personalization
[13]	Remote monitoring and data analytics	Sending alert signals	Heart rate, blood pressure, body temperature	Mobile and web interfaces for real-time monitoring of cardiac data
[14]	Wearable heat stroke detection device (WHDD)	Data from people exercising and fuzzy logic to assess the levels of stroke risk	Heart rate, galvanic skin response, body temperature	Long-rage data transmission and low power consumption

Table 2. Comparative analysis of features of healthcare systems.

Ref.	Wireless Sensors	Multimedia Sensors	Human Activity Recognition	Anomaly Detection	Fall Detection
[8]	Yes	No	Yes	No	No
[9]	Yes	Yes	No	No	Yes
[10]	Yes	No	Yes	Yes	Yes
[21]	Yes	Yes	No	Yes	Yes
[14]	Yes	No	No	Yes	No
[12]	Yes	No	Yes	Yes	Yes
[13]	Yes	No	No	Yes	Yes
[22]	Yes	No	No	Yes	No
Proposed	Yes	Yes	Yes	Yes	Yes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

MDPI and ACS Style

Gupta, A.; Sawhney, S.; Ahmed, S. CareTaker.ai—A Smart Health-Monitoring and Caretaker-Assistant System for Elder Healthcare. Eng. Proc. 2024, 78, 7. https://doi.org/10.3390/engproc2024078007

AMA Style

Gupta A, Sawhney S, Ahmed S. CareTaker.ai—A Smart Health-Monitoring and Caretaker-Assistant System for Elder Healthcare. Engineering Proceedings. 2024; 78(1):7. https://doi.org/10.3390/engproc2024078007

Chicago/Turabian Style

Gupta, Ankur, Sahil Sawhney, and Suhaib Ahmed. 2024. "CareTaker.ai—A Smart Health-Monitoring and Caretaker-Assistant System for Elder Healthcare" Engineering Proceedings 78, no. 1: 7. https://doi.org/10.3390/engproc2024078007

APA Style

Gupta, A., Sawhney, S., & Ahmed, S. (2024). CareTaker.ai—A Smart Health-Monitoring and Caretaker-Assistant System for Elder Healthcare. Engineering Proceedings, 78(1), 7. https://doi.org/10.3390/engproc2024078007

Article Menu

CareTaker.ai—A Smart Health-Monitoring and Caretaker-Assistant System for Elder Healthcare^†

Abstract

1. Introduction

2. Literature Review