A Conceptual IoT-Based Elderly Healthcare Monitoring Framework with Voice-Assisted Safety Verification

A Conceptual IoT-Based Elderly Healthcare Monitoring Framework with Voice-Assisted Safety Verification

Roopika. D, Meganadh. CH, Manikanta.M, Ajay.T,

student, BTech IoT 4th Year, Holy Mary Inst. Of Tech. And Science, Hyderabad, TS, India,

Aruna jyothi. P , Dr.V. Kotireddy. G, Dr. Venkataramana. B

Assoc. prof, CSE, holy mary inst. Of tech. and science, Hyderabad, TS, india,

Abstract: The increasing elderly population has created a growing demand for continuous, reliable, and non-intrusive healthcare monitoring solutions, particularly for individuals living independently. Conventional healthcare systems often lack real-time monitoring capabilities and timely emergency response mechanisms. This paper presents a conceptual IoT-based healthcare monitoring framework for elderly safety, integrated with machine learning techniques and a self-activating voice-assisted verification mechanism. The proposed framework conceptually employs wearable and ambient sensors to continuously monitor vital physiological parameters and movement patterns. Machine learning models are incorporated at the design level to support early identification of abnormal health conditions and potential fall events. A key contribution of this conceptual study is the inclusion of an automated voice interaction module that verifies the elderly users condition following a detected anomaly and initiates emergency alerts in the absence of user response. The paper focuses on system architecture, functional workflow, and design considerations, serving as a foundation for future implementation and experimental validation of intelligent elderly healthcare systems.

KEYWORDS : IoT-based Healthcare, Elderly Monitoring, Conceptual Framework, Machine Learning, Fall Detection, Voice- Assisted Safety, Smart Healthcare Systems

1. INTRODUCTION
   1. Background and Motivation

      The global rise in life expectancy has led to a significant increase in the elderly population, creating new challenges for healthcare systems worldwide. Elderly individuals often experience age-related health issues such as reduced mobility, chronic diseases, and a higher risk of falls, which require continuous monitoring and timely medical intervention. However, traditional healthcare approaches mainly rely on periodic hospital visits and manual supervision, making them insufficient for real-time health observation and rapid emergency response [1], [3]. These limitations motivate the development of intelligent healthcare solutions capable of providing continuous, automated, and remote monitoring for elderly individuals living independently.

   2. Importance of Elderly Healthcare Monitoring

      Continuous healthcare monitoring is essential for improving the safety and quality of life of elderly individuals, particularly those living alone. Falls and sudden physiological abnormalities are among the leading causes of serious injury and hospitalization in older adults [5]. Early detection of vital sign irregularities, such as abnormal heart rate, reduced oxygen saturation, or unusual movement patterns, enables timely preventive measures and emergency response [4]. Effective monitoring systems also reduce the burden on caregivers and healthcare facilities by minimizing unnecessary hospital visits and enabling home-based care solutions [2], [8].

   3. Role of IoT and Intelligent Systems

      The Internet of Things (IoT) plays a critical role in modern healthcare by enabling interconnected sensors and devices to continuously collect and transmit health-related data in real time [1]. Wearable and ambient IoT sensors allow non-intrusive monitoring of physiological parameters and daily activities, making them suitable for elderly users [4]. When integrated with intelligent systems such as machine learning and data analytics, IoT-based healthcare platforms can conceptually support automated anomaly detection, activity recognition, and early risk identification [7], [10]. Furthermore, the incorporation of natural language processing and voice-assisted interaction enhances system usability by enabling hands-free communication and emergency verification, which is particularly beneficial for elderly individuals during critical

      situations [9]. The convergence of IoT, intelligent algorithms, and voice-based interfaces therefore represents a promising direction for the design of advanced elderly healthcare monitoring frameworks

2. PROBLEM STATEMENT AND RESEARCH MOTIVATION
   1. Challenges Faced by Elderly People Living Alone

      Elderly individuals living independently are exposed to a range of physical, medical, and social challenges that increase their vulnerability to health emergencies. Age-related conditions such as reduced mobility, balance disorders, cardiovascular diseases, and respiratory issues significantly elevate the risk of falls and sudden medical complications [5]. In many cases, elderly individuals may be unable to seek immediate help due to physical limitations, unconsciousness, or panic during emergencies. Social isolation further aggravates these risks, as delayed assistance can lead to severe health consequences. These challenges highlight the need for continuous, automated healthcare monitoring systems that can function without constant human supervision [3].

   2. Limitations of Existing Healthcare Monitoring Systems

      Although several healthcare monitoring solutions have been proposed in recent years, many existing systems exhibit critical limitations. Conventional healthcare practices primarily rely on periodic clinical assessments, which fail to capture real-time health variations and emergency events occurring between visits [1]. Some IoT-based monitoring systems depend heavily on manual user input or wearable panic buttons, which may not be feasible during fall incidents or sudden health deterioration [2]. Additionally, many systems focus solely on sensor- based alerts without incorporating intelligent verification mechanisms, leading to false alarms and reduced system reliability [5]. Concerns related to data privacy, system usability, and the technological complexity of existing solutions also limit their adoption among elderly users [10].

   3. Need for a Voice-Assisted Verification Mechanism

      To address the shortcomings of existing monitoring approaches, there is a clear need for an intuitive and reliable verification mechanism that can confirm the elderly users condition following a detected anomaly. Voice-assisted interaction offers a natural and hands-free communication channel, enabling elderly users to respond to system prompts without requiring physical effort or technical expertise [9]. By integrating a voice-based verification step, the system can distinguish between actual emergencies and non-critical events, thereby reducing false alerts and unnecessary caregiver intervention. In situations where the user is unable to respond, the absence of a voice response itself can serve as an indicator of a critical condition, prompting automatic emergency notifications. This motivates the inclusion of a self-activating voice- assisted verification mechanism as a key component in intelligent elderly healthcare monitoring frameworks.

3. OBJECTIVES AND SCOPE OF THE STUDY
   1. Main Objectives of the Conceptual Framework

      The primary objective of this study is to propose a conceptual IoT-basd healthcare monitoring framework aimed at enhancing the safety and well-being of elderly individuals living independently. The framework is designed to integrate wearable and ambient sensors with intelligent data processing techniques to support continuous observation of physiological parameters and movement patterns. A key objective is to conceptually incorporate machine learning models capable of identifying abnormal health conditions and potential fall events based on multi- dimensional sensor data. Additionally, the framework aims to introduce a self-activating voice-assisted verification mechanism that interacts with the elderly user following detected anomalies, enabling confirmation of their condition and facilitating timely emergency response.

   2. Scope and Boundaries of the Proposed Work

      The scope of this study is limited to the design and conceptualization of a smart elderly healthcare monitoring framework. It focuses on outlining system architecture, functional components, data flow, and interaction mechanisms between sensors, intelligent processing modules, and user interfaces. The proposed framework considers commonly used physiological and motion sensors, cloud-based data management, and intelligent decision-support components. However, detailed hardware implementation, real-time deployment, system optimization, and large- scale performance evaluation are outside the scope of the present work. The framework is intended to be flexible and adaptable to different hardware platforms, communication technologies, and intelligent algorithms in future implementations.

   3. Clarification of Conceptual Design

      This study is presented strictly as a conceptual design paper. No physical hardware prototype, real-world data collection, or experimental validation has been conducted at this stage. All machine learning models, voice-assisted interaction mechanisms, and system workflows discussed in this paper are proposed at a design and planning level. Quantitative performance evaluation, accuracy analysis, and real-world testing are identified as future research directions. This clarification ensures transparency and positions the work as a foundational framework for subsequent implementation and experimental studies in intelligent elderly healthcare monitoring.

4. LITERATURE REVIEW
   1. IoT-Based Elderly Healthcare Monitoring Systems

      Recent research highlights the significant role of Internet of Things (IoT) technologies in enabling continuous healthcare monitoring for elderly individuals. IoT-based healthcare systems typically employ wearable sensors and ambient devices to collect physiological parameters such as heart rate, body temperature, oxygen saturation, and activity levels, allowing remote observation of health conditions in real time [1], [4]. Several studies emphasize that such systems improve emergency response time and reduce the need for frequent hospital visits, thereby supporting independent living for elderly users [3], [8].

      However, many existing IoT-based monitoring solutions primarily focus on data collection and alert generation without incorporating intelligent verification mechanisms. As a result, these systems often suffer from false alarms and limited adaptability to individual behavioral patterns, which can reduce user trust and system effectiveness [5]. Additionally, challenges related to data security, interoperability, and user acceptance remain open issues in large-scale IoT healthcare deployments [2].

   2. Machine Learning in Health Anomaly and Fall Detection

      Machine learning techniques have been increasingly applied in healthcare monitoring systems to enable automated analysis of sensor data and early detection of abnormal health conditions. Supervised learning models such as Support Vector Machines, Random Forests, and decision tree-based approaches have been widely explored for activity recognition and fall detection using wearable sensor data [5], [7]. These models are capable of identifying complex patterns in multi-dimensional data, making them suitable for analyzing physiological and motion-related signals in elderly healthcare applications.

      Deep learning methods have also been investigated for sensor-based activity recognition due to their ability to model temporal and non-linear relationships within data streams [7]. Despite these advancements, many studies rely on controlled datasets and laboratory environments, which may not accurately represent real-world elderly behavior. Furthermore, several existing systems trigger alerts solely based on sensor classification results, without incorporating contextual or user-driven verification, leading to increased false positive rates and unnecessary emergency notifications [10].

   3. Voice-Assisted and NLP-Based Healthcare Systems

      Voice-assisted systems and natural language processing (NLP) technologies have gained attention in smart healthcare applications due to their potential to provide intuitive, hands-free interaction. Voice-enabled assistants have been explored for medication reminders, health status inquiries, and patient support, particularly for elderly users with limited technical proficiency [9]. These systems improve accessibility and user engagement by allowing spoken communication instead of manual input.

      In the context of emergency healthcare monitoring, voice-based interaction has been proposed as a supplementary mechanism to sensor-based detection, enabling confirmation of the users condition after a suspected fall or health anomaly. However, existing voice-assisted healthcare solutions are often implemented as standalone applications or require explicit user activation, limiting their effectiveness during critical situations where the user may be unable to initiate interaction [9]. Integration of automated voice verification directly into IoT-based monitoring frameworks remains limited in current literature.

   4. Research Gap Identification

      From the reviewed literature, it is evident that IoT-based healthcare monitoring systems and machine learning-based anomaly detection techniques have been extensively studied. However, most existing solutions focus either on sensor-based monitoring or on intelligent data analysis in isolation. There is a notable lack of integrated frameworks that combine continuous IoT-based sensing, intelligent anomaly detection, and automated voice-assisted verification within a unified system architecture.

      Furthermore, existing systems often lack mechanisms to differentiate between true emergencies and non-critical events through user interaction, leading to false alerts and reduced system reliability. The absence of a self-activating voice-assisted verification step following anomaly detection represents a significant research gap. This study addresses this gap by proposing a conceptual framework that integrates IoT-based sensing, machine learning-driven anomaly detection, and automated voice interaction to enhance the safety and reliability of elderly healthcare monitoring systems.

5. PROPOSED CONCEPTUAL FRAMEWORK
   1. Overall System Architecture

      The proposed conceptual framework is designed as a multi-layered IoT-based healthcare monitoring system aimed at ensuring continuous observation and safety of elderly individuals living independently. The architecture integrates wearable and ambient sensing devices with intelligent data processing, cloud-based services, and user interaction modules. Sensor-generated data are collected from the elderly user and transmitted through a gateway device to cloud infrastructure for storage and analysis. Machine learning models are conceptually applied to identify abnormal health patterns and potential fall events. Upon detection of an anomaly, a voice-assisted verificaton mechanism is activated to confirm the users condition and initiate appropriate alerts. The layered architecture enables modularity, scalability, and flexibility, allowing future implementation using different hardware platforms and intelligent algorithms.

      Fig. 1. Conceptual architecture of the proposed IoT-based elderly healthcare monitoring framework with voice-assisted safety verification.

   2. Sensing Layer (Wearable and Ambient Sensors)

      The sensing layer forms the foundation of the proposed framework and is responsible for continuous data acquisition. Wearable sensors are conceptually employed to monitor physiological parameters such as heart rate, oxygen saturation, body temperature, and movement intensity. Motion sensors, including accelerometers and gyroscopes, are considered for capturing body posture and detecting sudden movements associated with falls. In addition to wearable devices, ambient sensors placed within the living environment can provide contextual information related to activity levels and surroundings. The use of non-intrusive and lightweight sensors is emphasized to ensure user comfort and long-term usability for elderly individuals.

   3. Communication and Gateway Layer

      The communication and gateway layer facilitates reliable data transmission between the sensing devices and the cloud infrastructure. A local gateway device is conceptually positioned to aggregate sensor data and manage communication protocols. Short-range wireless technologies such as Bluetooth Low Energy and Wi-Fi are considered for sensor-to-gateway communication, while internet-based protocols support data transmission to remote servers. The gateway layer also serves as an interface for integrating voice input and output components, enabling local interaction and preliminary data handling before cloud transmission.

   4. Cloud and Data Processing Layer

      The cloud and data processing layer acts as the central component for data storage, management, and intelligent analysis. Sensor data received from the gateway are conceptually stored in a secure cloud environment to maintain historical health records. Machine learning models are integrated at the design level to analyze incoming data streams and identify abnormal physiological patterns or movement behaviors. This layer supports real-time monitoring as well as long-term trend analysis, enabling proactive health assessment. Cloud-based processing allows scalability and facilitates future enhancements, including advanced analytics and integration with healthcare information systems.

   5. Application and Alert Layer

      The application and alert layer provides system interaction and information access for elderly users, caregivers, and healthcare professionals. User interfaces are conceptually designed to display health status, notifications, and alerts in a simple and accessible manner. In the event of a

      detected anomaly, the system activates a voice-assisted verification mechanism that interacts with the elderly user through spoken prompts. Based on the users response or lack thereof, alerts can be forwarded to caregivers or emergency contacts. This layer emphasizes ease of use, minimal user intervention, and timely communication during critical situations.

   6. Security and Privacy Considerations

      Security and privacy are critical design considerations in the proposed conceptual framework due to the sensitive nature of healthcare data. The framework conceptually incorporates secure data transmission mechanisms, user authentication, and controlled access to health information. Data encryption during transmission and storage is considered to protect against unauthorized access. Privacy-preserving design principles are emphasized to ensure that personal health data are handled responsibly and in compliance with healthcare data protection standards. These considerations are essential for building user trust and supporting real-world adoption in future implementations.

      Fig. 2. Layered representation of the proposed conceptual framework for elderly healthcare monitoring.

6. CONCEPTUAL METHODOLOGY
   1. Health Data Acquisition and Feature Representation

      In the proposed conceptual framework, health-related data are acquired through wearable and ambient sensors designed to operate continuously and non-intrusively. Physiological parameters such as heart rate, oxygen saturation, and body temperature are conceptually monitored to capture the health status of elderly individuals. Motion-related data, including acceleration and movement intensity, are considered for analyzing physical activity and detecting unusual movement patterns.

      The collected sensor readings are organized into structured feature representations suitable for intelligent analysis. Each data instance is conceptually represented as a feature vector comprising physiological and motion attributes. This representation enables unified processing of heterogeneous sensor data and facilitates subsequent analysis by machine learning models. Feature normalization and temporal aggregation are considered at the design level to handle variations across users and time periods.

      Fig. 3. Conceptual health data acquisition and feature representation process.

   2. Conceptual Machine Learning Framework for Anomaly Detection

      Machine learning techniques are conceptually integrated into the framework to support automated identification of abnormal health conditions and potential fall events. Supervised classification models, including decision tree-based approaches, ensemble learning methods, and support vector machines, are considered due to their suitability for multi-dimensional health data analysis. These models are conceptually designed to categorize health states into normal, warning, and critical conditions based on deviations from learned patterns.

      The framework emphasizes adaptability by allowing machine learning models to be trained using real-world elderly healthcare datasets in future implementations. At the conceptual level, the models are designed to process both real-time sensor inputs and historical data to support anomaly detection and trend analysis. This intelligent processing aims to reduce false alarms and enhance decision reliability by considering multiple physiological and movement indicators simultaneously.

      Fig. 4. Conceptual machine learning-based anomaly detection workflow.

   3. Conceptual Voice-Assisted Safety Verification Mechanism

      The proposed framework incorporates a voice-assisted safety verification mechanism as a key design component to improve emergency response reliability. Upon detection of a potential anomaly or fall event by the intelligent analysis module, the system conceptually activates an automated voice interaction with the elderly user. The voice module delivers simple spoken prompts to inquire about the users condition and request confirmation of safety.

      Natural language processing techniques are conceptually employed to interpret user responses and identify emergency-related intents. In cases where the user confirms safety, the system can suppress unnecessary alerts. Conversely, if the user requests assistance or fails to respond within a predefined time window, the absence of a response is treated as an indicator of a critical condition, triggering emergency notifications to caregivers or medical services. This voice-assisted verification approach enhances system reliability while maintaining ease of use for elderly individuals.

      Fig. 5. Conceptual workflow of the voice-assisted safety verification mechanism.

7. DISCUSSION
   1. Expected Advantages of the Proposed Framework

      The proposed conceptual framewor offers several advantages over conventional elderly healthcare monitoring approaches. By integrating continuous IoT-based sensing with intelligent data analysis, the framework supports proactive identification of potential health risks rather than relying solely on reactive emergency alerts. The inclusion of a self-activating voice-assisted verification mechanism enhances system reliability by enabling direct user interaction following detected anomalies. This approach helps differentiate between actual emergencies and non-critical events, thereby reducing unnecessary caregiver intervention and alert fatigue. Additionally, the modular and layered design of the framework promotes scalability and flexibility, allowing future extensions and adaptation to diverse healthcare scenarios.

   2. Practical Feasibility and Design Considerations

      From a practical perspective, the proposed framework is designed to be feasible using currently available IoT technologies and cloud-based services. Wearable sensors, wireless communication protocols, and voice interaction components are widely accessible and have been successfully applied in related healthcare applications. Design considerations such as user comfort, ease of use, and minimal manual interaction are emphasized to ensure suitability for elderly users. The framework also considers energy efficiency, reliable connectivity, and data security as important factors for real-world deployment. While the current study focuses on conceptual design, these considerations provide a realistic foundation for future implementation and system optimization.

   3. Comparison with Existing Approaches

      Compared to existing elderly healthcare monitoring systems, the proposed framework adopts a more integrated and user-centric approach. Many current solutions primarily rely on sensor-based detection and automated alerts, often without incorporating user verification mechanisms. In contrast, the proposed framework conceptually combines IoT-based sensing, machine learning-driven anomaly detection, and automated voice- assisted interaction within a unified architecture. This integration addresses limitations related to false alarms and limited user engagement observed in earlier systems. Furthermore, the framework emphasizes adaptability and extensibility, enabling future incorporation of advanced analytics, multi-language voice support, and integration with healthcare service providers. While existing approaches provide valuable foundations, the proposed conceptual framework aims to enhance reliability, usability, and overall system effectiveness through its holistic design.

      Fig. 6. Use-case scenario illustrating elderly fall detection and emergency response.

8. LIMITATIONS OF THE CONCEPTUAL STUDY
   1. Absence of Real-World Implementation

      A primary limitation of this study is the absence of a physical hardware prototype or real-world system deployment. The proposed framework is presented at a conceptual and architectural level, and no practical implementation has been carried out to evaluate system performance under real operating conditions. Consequently, factors such as real-time responsiveness, system robustness, and user interaction effectiveness have not been empirically assessed. This limitation is inherent to the conceptual nature of the study and is intended to be addressed in future implementation-focused research.

   2. Dataset and Validation Limitations

      The proposed machine learning and anomaly detection components are described conceptually and have not been trained or validated using real-world elderly healthcare datasets. As a result, the effectiveness of the proposed intelligent analysis in accurately detecting health anomalies and fall events cannot be quantitatively verified within the scope of this study. Variations in individual health conditions, movement behavior, and environmental factors are not empirically captured at this stage. Comprehensive dataset collection, model training, and validation are therefore identified as essential future research tasks.

   3. Hardware and Deployment Challenges

      Another limitation relates to potential hardware and deployment challenges that are not explored in depth in this conceptual work. Practical issues such as sensor calibration, battery life, device wearability, wireless connectivity reliability, and long-term system maintenance may impact real-world performance. Additionally, deployment in diverse living environments may introduce challenges related to network coverage, environmental noise affecting voice interaction, and user adaptation to new technologies. These practical considerations highlight the need for extensive testing and optimization during future system implementation.

9. FUTURE WORK
   1. Planned Hardware Implementation

      Future work will focus on the physical implementation of the proposed conceptual framework using suitable wearable and ambient sensing devices. This includes integrating physiological and motion sensors with a gateway device to enable real-time data acquisition and transmission. Hardware design considerations such as device miniaturization, energy efficiency, wearability, and long-term reliability will be addressed to ensure practical usability for elderly individuals. Real-world deployment in controlled and home environments will be explored to evaluate system functionality and user acceptance.

   2. Dataset Collection and Model Training

      Another important direction for future research involves the collection of real-world healthcare and movement datasets from elderly users. These datasets will be used to train and validate the proposed machine learning models for anomaly and fall detection. Emphasis will be placed on capturing diverse activity patterns, health conditions, and environmental variations to improve model generalization. Proper evaluation techniques, including cross-validation and performance benchmarking, will be applied to assess model effectiveness and reliability.

   3. Multi-Language Voice Support

      To enhance accessibility and inclusivity, future versions of the system will aim to support multi-language voice interaction, particularly regional and local languages. This extension is expected to improve usability for elderly individuals who may not be comfortable with English-based systems. Voice prompts and responses will be designed with clear pronunciation and adjustable speech speed to accommodate varying user needs. Incorporating multi-language natural language processing will further strengthen the systems applicability in diverse cultural and linguistic contexts.

   4. Mobile Application and Hospital Integration

      Future development will also include the design of a mobile application to provide caregivers and family members with real-time access to health data, alerts, and system status. The application will support health history visualization and notification management. Additionally, integration with healthcare providers and hospital information systems will be explored to enable faster emergency response and seamless communication during critical situations. Such integration aims to enhance coordination between home-based monitoring systems and formal healthcare services.

10. CONCLUSION

This paper presented a conceptual IoT-based healthcare monitoring framework aimed at enhancing the safety and well-being of elderly individuals living independently. The proposed framework integrates wearable and ambient sensing technologies with intelligent data processing and cloud-based services to support continuous monitoring of physiological parameters and movement patterns. By outlining a layerd system architecture and conceptual methodology, the study provides a structured foundation for designing intelligent elderly healthcare monitoring systems.

A key significance of the proposed framework lies in the integration of a self-activating voice-assisted safety verification mechanism. This feature enables intuitive and hands-free interaction with elderly users following detected anomalies, allowing the system to confirm user status and distinguish between critical emergencies and non-critical events. The voice-assisted approach enhances system reliability, reduces false alerts, and improves user engagement, particularly for elderly individuals with limited technical proficiency.

As a conceptual study, this work contributes to smart healthcare system design by identifying gaps in existing elderly monitoring solutions and proposing an integrated framework that combines IoT sensing, machine learning-based anomaly detection, and automated voice interaction. The framework serves as a reference model for future implementation, experimentation, and real-world deployment of intelligent elderly healthcare systems. By emphasizing usability, scalability, and safety, the proposed design offers valuable insights for researchers and practitioners working toward advanced and user-centric smart healthcare solutions.

REFERENCES

1. S. R. Islam, D. Kwak, M. H. Kabir, M. Hossain, and K. S. Kwak, The Internet of Things for Health Care: A Comprehensive Survey, IEEE Access, vol. 3,

   pp. 678708, 2015.

2. P. Gope and T. Hwang, BSN-Care: A Secure IoT-Based Modern Healthcare System Using Body Sensor Network, IEEE Sensors Journal, vol. 16, no. 5, pp. 13681376, 2016.
3. A. Botia, J. Villa, and D. Palma, Ambient Assisted Living System for In-Home Monitoring of Elderly People, Journal of Medical Systems, vol. 36, no. 3,

   pp. 14511461, 2012.

4. S. Patel, H. Park, P. Bonato, L. Chan, and M. Rodgers, A Review of Wearable Sensors and Systems with Application in Rehabilitation, Journal of NeuroEngineering and Rehabilitation, vol. 9, no. 21, 2012.
5. N. Pannurat, S. Thiemjarus, and E. Nantajeewarawat, Automatic Fall Monitoring: A Review, Sensors, vol. 14, no. 7, pp. 1290012936, 2014.
6. A. Reiss and D. Stricker, Introducing a New Benchmarked Dataset for Activity Monitoring, IEEE International Symposium on Wearable Computers, pp. 108109, 2012.
7. J. Wang, Y. Chen, S. Hao, X. Peng, and L. Hu, Deep Learning for Sensor-Based Activity Recognition: A Survey, Pattern Recognition Letters, vol. 119,

   pp. 311, 2019.

8. M. R. Ruman, S. M. Salaken, and A. Rahman, IoT-Based Remote Health Monitoring System, IEEE International Conference on Power, Computing and Communication Technologies (CPCT), 2020.
9. T. Zhou, H. Wang, and Y. Li, Natural Language Processing Techniques for Smart Healthcare Applications, arXiv preprint arXiv:2103.12098, 2021.