1 Introduction

The elderly in smart home environments always require more remote medical support; they also need to visit doctors frequently [1]. The implementation of wearable sensors helps to design suitable healthcare systems for the elderly in smart homes [3]. It has been identified that the human and machine interface increased vividly [6]. In this way, smart sensors are used for monitoring vital healthcare parameters of the elderly in smart homes [12]. Elderly healthcare is maintained in the context of home-oriented health monitoring systems, monitoring of the elderly, and remote medical [19]. Elderly healthcare monitoring is the systematic and continuous tracking, assessment, and management of the health and well-being of elderly individuals.

Furthermore, smart sensor technology is employed to offer assistance and alleviate discomfort for the elders. Thus, it is essential to design robust sensor-based healthcare monitoring and tracking [71]. Therefore, the research aimed to monitor and track vital healthcare parameters for maintaining elderly care [11]. Also, monitor and track the situation of elderly people in the smart home environment. This can be done through wearable sensor technology accurately attached to the elderly by using various intelligent sensors that monitor and analyze their activities such as heartbeat, etcetera [15]. The components of the system are sensing data storage and data communication [7]. These components are important for wearable sensor technology for elderly care [67].

To define the issue of certain research domains, wearable sensor technology is implemented by remote monitoring techniques for people [22]. The particular technology is applied in the domain of elderly care to monitor and track healthcare parameters [90]. The limitation of the technology is identified as inaccurate monitoring of various healthcare parameters [24]. Other limitations of wearable sensors include the intrusive nature and potential user discomfort. In practice, the choice of wearable sensors depends on the specific application and user comfort considerations. The proposed taxonomy aims to implement wearable sensors for healthcare parameter monitoring [34]. Moreover, the smart sensors monitor and further track the vital parameters for maintaining elderly care [88].

Towards that, SDD (Sensing, Data storage, Data communication) taxonomy is used so that wearable technology and monitoring techniques can be accurately combined in elderly care [54]. In this taxonomy, we followed a similar methodology used in [91] and [92], which has been very useful and helpful in healthcare.

The research focuses on analyzing and evaluating the latest literature on wearable sensor technology and monitoring and tracking techniques used for monitoring and tracking various healthcare parameters from the elderly’s body [35]. Also, the feasible monitoring of healthcare parameters improved the wearable technology system towards providing robust and sophisticated services to the elderly living in smart home environments [62, 72]. Moreover, classifying various major components helps analyze and evaluate solutions to existing problems [37]. The detailed system is quite crucial in the elderly care domain. This is because remote monitoring techniques are more efficient than other healthcare techniques [56].

There is a limitation in the current solution, which could have provided real-time healthcare monitoring and tracking frequently [69]. The detailed model improved the problems identified in the existing solution by implementing a robust SDD taxonomy that helps regularly and frequently monitor and track elderly healthcare parameters in the smart home environment [57]. Further, SDD taxonomy helps in sensing various healthcare parameters, sensed data classification on the required criteria, and communication of various classified and analyzed data to the respective healthcare provider to achieve high accuracy and healthcare security for elderly people in the smart home environment [45].

The research project is structured into distinct sections. Section 2 commences with a literature review, where numerous papers are gathered, and relevant components are identified. Sections 3 and 4 also offer an advanced solution by utilizing the system components. Section 5 introduces a proposed model that enhances the state-of-the-art solution using algorithms and diagrams. Section 6 evaluates and validates the system, while Section 7 verifies the proposed system. Finally, Section 8 and 9 delve into discussing the research findings and presenting the conclusions, respectively.

2 Literature review

The other researchers have used fewer components in their taxonomy, such as one or two, but three central components are used in this research. Further, our proposed SDD taxonomy aims to help researchers improve healthcare monitoring of various healthcare parameters through wearable sensors. The research has focused on proposed techniques and algorithms that can be used to improve the accuracy of monitoring and tracking various healthcare parameters of the elderly in the smart home environment. The methods and tools suitable for the research domain can be deployed to improve the monitoring accuracy of healthcare parameters in the healthcare industry. The literature review section is divided into three major sections based on the SDD taxonomy: sensing various healthcare parameters using wearable sensors, storing sensed data using monitoring techniques, and communicating healthcare data.

Robot-integrated smart home (RiSH) has been proposed for improving the security of elderly people in smart homes and monitoring various healthcare parameters through wearable sensors [70]. They proposed a solution to resolve the existing problem by using RiSH, a robotics-based system that improved the safety of elderly people in smart homes. The result reveals that the overall accuracy of healthcare parameters detection has achieved 80% [13]. In the words of [4], accelerometer sensors are used for accurately sensing walking and movement-related motions. An improvement of 78.69% has been achieved by implementing the proposed machine learning algorithm in comparing the existing solution [76]. Also, the elderly walking is monitored to improve the healthcare of elderly people. This solution provides a special prototype that develops and reflects in the health care of the human. The Optical flow feedback convolution neural network for improving elderly posture recognition was proposed by [64]. The features that are identified in the system are machine learning models, motion detection, and recognition. Further, the results reveal that the current ratio has achieved 82.7% accuracy. Constrained Markov Decision Process (CMPD), Adaptive Learning Algorithm introduced towards optimizing the energy consumption in intra and beyond networks while communicating the monitored healthcare data, proposed by [1]. The intelligent algorithm has proven to achieve a 100% throughput improvement in various power consumption budgets [55]. The current solution is much more valuable because of its advancement than the other available solutions and has achieved all the promised results. BAN logic model and mutual authentication have been detailed for improving patient health and the precision medical field [83]. The current solution is valuable because it solves two major problems. First, it increases the productivity of the medical field, and second, it increases the network's lifespan by minimizing energy consumption [58]. The proposed solution has increased the network's life and the production of healthcare monitoring.

An established machine learning-based accelerometer used to monitor the walking parameters for improving the healthcare of elderly people was introduced by [63]. Further, inertial motion units and accelerometer signals are used for communicating the accelerometer data [61]. It provided vital information about the elderly walking, i.e., age and surface information. Moreover, the precision achieved 95.2%. Machine learning algorithm-based robust wearable sensors to monitor the healthcare parameters for improving the health of elderly people in the smart home environment was proposed by [38]. Towards that, robust biomedical sensors and wearable inertial sensors are equipped for domestic monitoring [36]. Even though data storage and high operational costs remain issues in the proposed system, the equipped sensors achieved an accuracy of (mean ± std = 0.99 ± 0.01). The proposed activity testing algorithm aimed to present an activity testing algorithm towards improving the healthcare of the elderly in smart homes with feasible activity prediction [10]. Recurrent neural networks are used for encoding sequential time information [59]. The detailed solution has achieved a mean accuracy of 99.52% during the sensor-based features based on Recurrent Neural Network (RNN). Other machine learning and deep learning algorithms were proposed to monitor various activities performed by the elderly in smart homes [32]. The convolutional neural network-based (CNN) algorithms have improved the classification accuracy of activity recognition. The CNN layers achieved a classification accuracy of 90.9%, even with the limited number of sensors [53].

Big data analytics have been detailed in managing healthcare data using Internet of Things (IoT) devices [5, 95]. Further, the sleep disorder factors are monitored through robust wearable sensors. The cloud layer accurately manages, analyzes, and stores the monitored data [8]. The proposed system monitored the air quality index prediction with 93.3% accuracy. The engine based on the formal concept analysis was proposed to provide real-time cognitive assistance [66]. The formal concept analysis proposed analysis system has been used to improve activity recognition accuracy. The results reveal that the data repetition error has been 100% detected [2]. Heterogeneous features are selected using the J48 decision tree model. Snowball sampling and Partial Least Square SEM (PLS-SEM) were proposed for analyzing the healthcare data in the smart home environment [31]. The system has improved while analyzing the healthcare data with a rate of ~ 89.67% from the current solution. Smart homes are highly advanced and provide cognitive assistance to elderly people. The Optimal Relay Placement Algorithm (ORPA) was used to increase the radio signal's obstacle-overcoming capability and find the best location to place the relays in an environment [50]. This research proposed a capable greedy algorithm ORPA that was used to calculate the best location for relay placement. Further, an improvement of 79.89% has been achieved in the current solution [16].

Established Smartphone applications and Wearable Sensors for Smart Healthcare Monitoring Systems (SW-SHMS) towards dynamic access control to healthcare data and provide better treatment to the patient [87]. The research helped in overcoming the security flaws by proposing a mutual trust-based model in the healthcare environment. The centric positioning techniques have achieved localization accuracy [18].

The Support Vector Machine-based classifier was proposed for managing monitored healthcare data [68]. The nodes above a threshold are chosen to be cluster nodes, and by using body sensors, patient health-related monitoring processes are performed [20]. Further, the energy efficiency of the SVM-based classifier has increased, along with the accuracy of 89.65% compared to the existing data management system [77]. The mobile phone-based algorithm was proposed for improving the gait speed estimation accuracy using machine learning algorithm-based wearable sensors [44, 78]. BioStampRC sensors were used for posterior directions [21]. Furthermore, the BioStampRC sensors accurately estimated the gait speed at 98.7%. The machine learning algorithm-based wearable sensors were introduced for identifying the diagnosis of the elderly. The feasibility of diagnosis monitoring has been achieved through the robust wearable sensors worn by the elderly living in the smart home environment [42, 96, 101]. The monitored data is further communicated to the healthcare provider in the context of the healthcare industry [25]. The belt-worn IMU has measured the angular velocity and acceleration with an accuracy of 98.01%. Activity recognition and lightweight algorithms were introduced to achieve high accuracy while recognizing elderly activities [52]. The codebook is used for feature representation, and sequences are converted into a feature representation with an accuracy of 89.89% [26]. Although latency has been identified in the proposed activity monitoring of the elderly in the context of the healthcare industry, the sensors are adequately used for measuring acceleration forces.

A depth sensor-based approach is proposed for monitoring and detecting the healthcare of elderly people [33]. The fall detection algorithm is used to detect the person through the robust Kinect sensors. The movement of the elderly is detected, monitored, and evaluated through the Y or Z coordinates [28]. The proposed solution gains an accuracy of 86.83%. Enhanced the health and living in their homes with the Unified Theory of Acceptance and Use of Technology (UTAUT) Framework to provide better management and online healthcare treatment [30]. Smart homes use smart devices to provide protection and healthy living for elderly people. It provides the security of elderly people has been improved by 81.4%. According to [13], the human body's activity is detected and enhanced with the help of a smartwatch or other devices. The Forward–Backward algorithm, the Diffie-Hellman algorithm, is used for solving the issue of the sitting-related problem by implementing sensor-based devices such as Smartwatches, etc. An improvement of 89.56% has been identified compared with the existing solution [41]. Smart intelligent and multimodal systems were implemented to solve the security issue of smart home healthcare systems. The controlling home appliances used radio frequency-based home automation to control all home appliances [39, 97, 104]. Wireless Bluetooth technology provides remote access from PC control appliances, which helps in feasible data communication [43]. Advanced processing and machine learning algorithms have been introduced to accurately process the monitored data [47, 98, 100]. The W3C semantic sensors are ontology and are used for observation in smart IoT environments to complete some tasks. The proposed algorithm achieved an accuracy of 89.76%.

Wearable sensors are feasible for tracking various healthcare parameters [17, 102]. More precisely, the proposed system aimed at implementing wearable sensors to monitor sleep timing to predict migraine attacks [40] better. Further, Quadratic discriminant analysis (QDA) achieved 84.1% accuracy than the existing solution [29]. Moreover, the classifiers achieved an accuracy of 91.2%, a sensitivity of 99.6%, and a specificity of 90.0%. Further, a Long Short-Term Memory (LSTM) network and wearable sensor were presented to provide feasible human activity recognition [14]. The proposed deep convolutional network and machine learning application help to monitor the data through robust wearable sensors [60]. It provides the testing time reduced by 38% through the backpropagation algorithm. Moreover, the results improved by 8% than the current solution classifying running speed conditions using a single wearable sensor in the context of elderly in smart homes [80]. The wearable sensors and feature extraction algorithm are used for accurate monitoring of speed conditions [74]. Moreover, MATLAB software and wearable sensors are equipped for monitoring the healthcare parameters of the elderly body [48]. The segmentation method achieved five-stride (97.49 (± 4.57) %) with the greatest classification accuracy.

Human activity monitoring data is accurately assessed through the predicted mean vote (PMV) index [23]. They introduced personal thermal comfort assessment using optimization techniques to improve human fall detection accuracy. It provides the sensors' accuracy of ± 2% for humidity and ± 0.5 ◦C for air temperature [27]. A model presented SmartFall towards improving the fall detection of the elderly in smart homes [46, 75]. They used SmartFall to identify the falls of the elderly in smart homes and communication of monitored data through wearable technology-based smartwatches. The real-time fallen data is detected accurately [49]. The missed fall is presented as false positives (FPs) that reduce the chances of time consumption in the system [51]. The improvement can be achieved through Nexus 5X smart phone with 1.8 GHz hexa-core processor, which is used as a data communication tool in the detailed solution [85]. A practical solution WISE system for adequate healthcare monitoring of elderly people was introduced by [9]. The proposed solution Wearable IoTcloud-baSed hEalth monitoring system (WISE), improved the performance of healthcare monitoring through the accelerometer sensors. The smartphone-based accelerometer sensor provided an accelerometer rate with an accuracy of 98.6%. A one-class support vector machine to define typical movement patterns was introduced by [53]. The wearable sensors have been used for monitoring the accelerometer rate and gyroscopic rate of the object through robust sensors. iNEMO inertial module was used as the inertial sensor to measure the velocity of the object [84]. The research has achieved an overall accuracy of 90%. However, future work shall focus on examining the impact of gait data [86].

The limitation of the range of sensors remains in the system [70]. The limitation of computational sources is identified in the research, but the improvement can be seen in the future work of the research [4]. The system has the limitation of inaccurate moving object identification that causes errors [64, 65]. The system has identified that the process needs extreme expert supervision for smooth completion, but the Lagrangian approach facilitates the transformation of the problem to provide an easy solution [1]. A minor limitation has occurred due to the involvement of many external factors. The limitation of the system is that features were not extracted accurately [63]. Data storage remains an issue, and high computational costs are the drawbacks of this system [38]. Their system has identified that gyroscopic data was not collected with precision and accuracy due to a lack of adaptability [10]. The proposed solution has identified the lack of identity approximation, but the improvement can be achieved by applying a sparse auto-encoder in the optimum feature location [32]. In [93], they analyze the shared requirements of elderly and disabled individuals, after which we assess a variety of IoT applications capable of delivering the necessary assistance. In [66], the system's limitation is that the features are not shared accurately. In [31], the major limitation identified in the proposed system is that the objects are not presented geometrically. The limitations of complexity and complicated processes are placed in the system [50]. The requirement of keen observation and expert supervision remains a major limitation in the research [87]. The limitation of high complexity in the monitoring processes is identified in the proposed solution [68]. Inadequate device location has been notified as the limitation of the proposed solution, but the noticeable improvement is identified through gait walking speed estimation during the over-ground ambulation [44]. Moreover, the temporal phases are not identified adequately due to the high error rate in the system [42]. However, latency has been identified in the proposed activity monitoring of the elderly in the context of the healthcare industry [52]. Nonetheless, the proposed sensor-based system does not accurately monitor the direction of movements [33]. Sometimes, technical issues occur in the system due to inaccurate sensing, but smart device performance is increased with the help of smart technologies using IoT [30, 99]. The system's limitation of inaccurate data communication remains, but sensors use Visual recognition-enabled devices to monitor body language [13]. In their system, data communication is found to be less secure due to inaccurate networking, but the Arduino prototyping board can be used to see which devices are working or not [39]. The limitation of limited interaction between humans and robots has been identified, but improvement could be seen through data processing services [47]. In [71], the system has identified that the inaccurate recognition of various healthcare parameters remains a limitation. The limitation in [14] is this system's motion detection difficulty. The limited storage capacity and inaccurate running condition monitoring have not been performed accurately [80]. The high latency was identified when extracting the sensed data [23]. The complexity was identified at the time of data extraction [46]. Limited memory, Storage capacity, and computing capacity are not performed accurately [9]. Limited sensor information and sample size prevented the work to accurate [53]. In [103], the aim was to help the elderly choose suitable physical activity and healthcare monitoring devices.

Towards monitoring and tracking various healthcare parameters from elderly people, monitoring techniques are used with the different methods described in the literature review section. It has been identified that literature number 2 of week 1 was the most accurate and relevant to the research domain. This literature includes the Hidden Markov Model (HMM) and the healthcare parameter monitoring technique to monitor elderly people's healthcare. The system consists of various components, i.e., the diffie-hellman algorithm, smartwatches, and binary images. In Fig. 1, the architecture for elderly assisted living in healthcare using wearable sensor technologies is provided. However, Fig. 2 shows the state-of-the-art solution.

Fig. 1
figure 1

Architecture for elderly assisted living in healthcare using wearable sensor technologies

Full size image
Fig. 2
figure 2

The State-of-art solution

Full size image

3 Proposed Framework

The current cutting-edge solution is hindered by issues such as unreliable monitoring, insufficient recognition of activities, inaccurate data synchronization, and imprecise data classification. In order to address these shortcomings, an enhanced solution has been developed, incorporating components that enhance the precision of healthcare parameter monitoring and tracking through the utilization of smart wearable sensor technology. Figure 3 depicts the proposed architectural framework and the new added features.

Fig. 3
figure 3

The proposed solution. Green rectangles show the added components

Full size image

4 Framework Components

The Sensing, Data storage, and Data communication (SDD) taxonomy was conducted to review current wearable sensor technologies and knowledge. In this taxonomy, we followed the same structure as [91] and [92]. Further, healthcare monitoring and tracking are evaluated in the background of elderly assisted living. Towards that, SDD taxonomy is conducted where sensing, data storage, and data communication help to sense and track the vital healthcare data for maintaining the health of the elderly living. Moreover, the evaluation, validation, and advancement of the system are going to be treated in this report.

In the report, 150 research articles have been reviewed and considered. Further, we identified that only 30 articles met the research domain requirements. There are some principles and standards that need to be focused on. In this research, wearable sensor technology is described in the context of elderly assisted living because it helps monitor and track vital healthcare parameters. Moreover, the research introduced a robust methodology to maintain the health of elderly people.

By going through the literature review, considerable knowledge and ideas were gathered. Consequently, to improve the remote monitoring and tracking of essential parameters of elderly people in the healthcare industry, comprehensive insight is required to be prepared. Three major facts that are needed to be considered in the research are as follows –

  • Provide the major parameters that assist the elderly in maintaining health and regular diagnosis.

  • What are the ways through which the relevant data can be collected and merged in this research?

  • How can tracking and monitoring healthcare parameters improve elderly assisted living?

Subsequently, considering these three facts, wearable sensor technologies are classified for healthcare monitoring and tracking vital information within the healthcare industry.

4.1 Taxonomy for Healthcare Monitoring and Tracking

Initially, the first factor used in the SDD taxonomy is sensing. Through the sensing or tracking stage, healthcare-related vital information is sensed through wearable sensor technologies. Moreover, Artificial Intelligence (AI) technologies are used for collecting physiological data. The sensed data is essential because it would help to provide elderly assisted living. The components and sub-components of sensing are presented in Table 1. In addition, the sensed data stored in the cloud, known as electronic healthcare data, is based on machine learning technology. This sensed healthcare data could be retrieved and communicated at any time. Once the data is stored, it can be further transmitted to the healthcare provider or elderly assistant. Hence, the feasible sensing and tracking of vital healthcare parameters can improve assisted elderly living. All three components of SDD taxonomy is further sub-classified along with their respective example or instances in Fig. 4. It can be further seen that these components are interrelated to each other. The instances in Table I underline the importance of particular factors in a certain domain.

Table 1 Main Attributes & Common Instances of the Sensing, Data Storage, and Data Communication
Full size table
Fig. 4
figure 4

The three factors of Monitoring and Tracking taxonomy (i.e., Sensing, Data Storage, and Data communication)

Full size image

In Fig. 4, all three major SDD taxonomy components are proposed. Further, all components interact with each other in a way to get a better understanding of the research domain. The remaining portion contains the evaluation, verification, and classification of the introduced SDD taxonomy. Further, these aspects are calculated in the classification of the system. More precisely, Fig. 4 represents the classes and sub-classes of the SDD taxonomy.

4.1.1 Sensing

This section contains two sub-classes of sensing main components: wearable sensors and cloud data. In addition, their functionality includes sensing, tracking, and monitoring. The classes and sub-classes are provided in Table 1 and Fig. 5 correlated to each other with their instances. Detailed data classes and sub-classes are measured towards addressing the requirement in the SDD taxonomy for feasible classification and specification of each type of data deployed in the research review.

Fig. 5
figure 5

Main attributes and instances of Sensing

Full size image

4.1.2 Data storage

In the way to manage and store the data, various data classifiers are used, which help to classify the healthcare data sensed through the wearable sensors. In the same way, wearable sensors and activity recognition are used for accurate storage analysis of the elderly monitoring results (Fig. 6). Support vector machine (SVM) and engine-based concept analysis are applied for further analysis.

Fig. 6
figure 6

Main attributes and instances of Data Storage

Full size image

4.1.3 Data communication

Once the data is sensed and stored, the healthcare data is communicated through classification, data analysis, and healthcare surveillance (Fig. 7). The healthcare provider and elderly can retrieve the data through the security claim. Figure 7 provides the flowchart of the data communication and its main classes and sub-classes.

Fig. 7
figure 7

Main attributes and instances of Data Communication

Full size image

4.2 Purpose of the Proposed Framework

4.2.1 Sensing

The aim of considering sensing as part of elderly assistance is to carry out what data needs to be sensed to improve elderly health. In other words, sensing, tracking, and monitoring are aimed in this section. Furthermore, the data is collected and communicated to the cloud environment to maintain sensed data security.

4.2.2 Data storage

Wireless sensor networks are used to manage and store the data. This section stores data such as activity recognition, various healthcare data, and wearable sensor-based data. Furthermore, the data is stored in the cloud using feasible wireless communication networks. All the sub-classes of data storage are mentioned in Fig. 3.

4.2.3 Data communication

The sensed data is communicated to respected healthcare providers using the lightweight algorithm and big data analyzers. The sub-classes are divided into classification, data analysis, and healthcare surveillance. In addition, the data is communicated to healthcare providers using a support vector machine-based classifier and convolution neural network.

5 Framework Classification

Around 130 results or journals have been evaluated and analyzed in the research. Of 130 results, 30 journals have been finalized to conduct the research in the given research domain. Further, for getting detailed and established information about the research, only journals published in 2018–2019 were considered. More precisely, Q1 and Q2 level journals as these journals are more precise and contain detailed and established information about the research. The journals are only collected on the basis of activity recognition and monitoring techniques. All 30 journals are classified and evaluated in the below section based on the proposed taxonomy and its useful components for constructing this research. Thus, the classification is prepared on 30 selected journals. In Table 2, various components are evaluated based on the proposed taxonomy and its components that are useful in obtaining detailed information about the research in the context of elderly care.

Table 2 Classification Of The Systems Based On The Proposed Sdd Taxonomy
Full size table

In Table 3: the classification and evaluation are done based on sensing components and various parameters that provide detailed information about the sensors that can be deployed for elderly care.

Table 3 Classification Of Systems On The Background Of Sensing Components
Full size table

In Table 4, the systems are classified in the background of various techniques and methods that are used for providing data storage to the data sensed and monitored through the various smart sensors.

Table 4 Classification Of The Systems Based On Data Storage Components
Full size table

Table 5 presents the evaluation of various parameters in the context of data communication to the research healthcare and security provider for improving and maintaining the elderly people care in the background of smart homes environment.

Table 5 Classification Of The Systems In The Context Of Data Communication Components
Full size table

5.1 Sensing

Various journals are aimed at the sensing component and utilization of this component in their research works. The sub-components, such as wearable sensors and clouds, are essential for the research.

  1. 1.

    Wearable sensor: In [5], they stated that the feasible healthcare tracking and monitoring of healthcare parameters are achieved by implementing an Ensemble machine learning algorithm within the system. Also, it helps monitor the healthcare parameters in a real-time environment and allows communication with the respective healthcare professionals.

  2. 2.

    Cloud: After sensing the healthcare parameters, the sensed data is Communicated and stored in the cloud environment. Towards that, an Optical flow feedback convolution neural network has been implemented for accurate healthcare data communication and storage [64].

5.2 Data storage

The data storage component is essential because it must be stored securely and efficiently before communicating the healthcare-sensed data. Towards that, the data storage component is sub-devices into two different sub-components, i.e., elderly monitoring and storage analysis.

  1. 1.

    Elderly monitoring: Elderly monitoring has been aimed in the research to identify the vital healthcare parameters for communicating with the healthcare provider. Big data analytics introduces health monitoring [5, 94]. Moreover, fall detection and human activity recognition are accurately performed through a One-class support vector machine (OCSVM) [53].

  2. 2.

    Storage analysis: After monitoring elderly healthcare parameters, the storage is accurately analyzed for managing extensive healthcare-related data. Towards that, Multi-dimensional Scaling and K-fold cross-validation type storage managers are used to manage the healthcare system data [13].

5.3 Data communication

In the previous two system components, healthcare data is tracked, monitored, and stored in a cloud environment to maintain healthcare data security and accessibility. Moreover, data communication is divided into three sub-components: classification, data analysis, and healthcare surveillance.

  1. 1.

    Classification: A Support Vector Machine-based classifier is introduced for the feasible classification of various healthcare data within the healthcare management system. In addition, classification is performed for accurate healthcare data segregation [31].

  2. 2.

    Data analysis: Rule-based filters are used for feasible healthcare data analysis. Data analysis is required because various types of vital healthcare parameters are sensed through the sensors. Thus, it needs to be analyzed on a regular basis for feasible healthcare data management [64].

  3. 3.

    Healthcare surveillance: In order to provide accurate healthcare surveillance to elderly people in the healthcare data management system, Sensor Nodes and AVISPA software are found to be the most reliable and precise. This is because it helps to improve healthcare data management accessibility; thus, healthcare providers and patients can access the data at any time [83].

Thus, the overall classification is performed to provide generic criteria for evaluating all 30 journals in the context of SDD taxonomy components. The healthcare data is accurately tracked and monitored using smart wearable sensors and evaluating sensing, data management, and data communication components.

6 System Components Validation And Evaluation

While determining various relevant factors of the wearable sensing techniques, the component that adds value to the system needs to be evaluated and validated. The value-added components are crucial for evaluating and validating the research domain's context. Meanwhile, all the publications used in the system use certain evaluation and validation models. For instance, most literature is aimed at wearable sensing technology for monitoring and tracking healthcare parameters in elderly care. This is because monitoring and tracking healthcare parameters is important; thus, the accuracy and efficiency of the monitoring need to be high. Table 6 lists all validation and evaluation of health care monitoring and tracking techniques with their mathematical Formula.

Table 6 Validation And Evaluation Of Health Care Monitoring And Tracking Techniques
Full size table

Many works of literature that are used in the research have contained limited or less information that makes the dataset incomplete. The uncompleted dataset can be stated as a single dataset. Moreover, the literature has not been verified for the complex and varied datasets. Thus, the single dataset validation has not been giving assurance for the system's performance on efficient levels due to the inability and inaccuracy to monitor and track the healthcare parameters of the elderly. The inaccurate healthcare monitoring leads to inadequate communication of healthcare parameters to healthcare professionals. Hence, the system must validate and evaluate the monitoring, tracking, and communication of data.

Monitoring and tracking sensors' performance is also important in evaluating and validating the system. Thus, the monitoring and tracking of various healthcare parameters need to be done on a regular basis to attain high accuracy. The sensing devices are required to be performed up to the prospect and provide the possible result to the system.

It has been identified that there are numerous literature journals where the evaluation of the system has not been performed through experiments and simulation. Further, the conceptual recommendation of the papers is stated through the numerical analysis of the model. Moreover, the research literature has not provided information on the background of real-time environments.

7 System verification

7.1 Description

The components used in the proposed system need to be evaluated due to their importance. In the same way, quantitative and qualitative methods are implemented in a mannered way toward evaluating the mechanisms. Moreover, the numerical analysis method is used to analyze the system. The proposed feasible system has solved the issue of inaccurate healthcare monitoring. In addition, the component existence can be compared with the proposed SDD taxonomy. The occurrence is presented in a percentage formation.

7.2 System identification

Towards evaluating the SDD taxonomy, a teat of overlap is conducted because it provides the occurrence of components and sub-components in the literature. The occurrence can be provided in the numerical value for more detailed information about the component occurrences. Moreover, sensing of healthcare parameters, monitoring, and tracking identified the component's time in the literature.

7.3 Completeness

While ensuring the system's completeness, evaluation of the major components and sub-components in the state-of-art. In this way, 30 journals were presented with the quality of Q1 and Q2 type journals. In this way, several journals have been aimed at healthcare parameter sensing and monitoring techniques in the context of smart home environments and elderly care. Figure 8 presents the percentage of component and sub-component occurrence with the proper demonstration. Table 7 lists all term frequencies used in the 30 publications.

Fig. 8
figure 8

Occurrence of the classes and components in the analysed literature, i.e., 90% is the occurrence rate of cloud components

Full size image
Table 7 Term Frequency For 30 Publications
Full size table

Figure 8 details the evaluation and analysis of various components and sub-components. The existing SDD taxonomy can be disciplined in the data classified and fog layer context. It has also been identified that the existence of these components is the least. Also, the components are the least overlapped in the system. More precisely, if the SDD taxonomy is not applicable, then the taxonomy shall be penalized in the context of the elderly healthcare environment.

8 Discussion

8.1 Justification

In this section, a discussion is going to be undertaken on which care components are to be discussed that are rarely discussed in the research literature. To provide the importance of certain components, instances are figured out from the literature connected to the components in the SDD (Sensing, Data Storage, and Data communication) taxonomy and its standards. The discussion section also highlights the value of evaluating various taxonomy components.

8.2 Sub-factor of each component

  1. 1.1.1.

    Sensing – Cloud: In the various literature that is reviewed in the research, most literature has discussed wearable sensor technology and detailed the sensors in their pieces of literature. Yet, the cloud has not been discussed in the research in the context of healthcare parameter tracking and monitoring. Cloud technology helps to communicate and monitor healthcare data with the respective medical professionals. Though the researchers have not discussed it, only two pieces of literature detailed the cloud.

Different kinds of sensing methods can be used for sensing various healthcare parameters from elderly people. Sensing, tracking, and monitoring methods are included in the literature. Yet, it has been noticed that cloud and its common instances were not included in the literature. Hence, the passage of these methods is not required and should not be considered. Nonetheless, the literature section discusses the components of Sensing, tracking, and monitoring.

  1. 2.2.2.

    Data storage – Storage analysis: It has been identified so far that the literature has not included storage analysis in their research. Even though it has been noticed that the data storage has been detailed by various researchers in their literature, data synchronization in the storage analysis is important for the research because it helps get information about the healthcare data that needs to be stored in the external storage sources. After all, only 1 article has detailed information about storage analysis, but it is to be considered that the storage analysis should be discussed more in the research as it helps to manipulate the healthcare data.

Various data storage processes are not detailed in the research so far, such as Support vector machine, engine-based concept analysis, etc. Moreover, these components have not been discussed and are not required to be because of their low weightage in the research domain. Yet, elderly monitoring and its components are discussed more in the research publication due to its high reliability.

  1. 3.3.3.

    Data communication – Classification: It can be concluded that almost every researcher has included data communication in their research and literature. Yet, the classification components have not been discussed in detail. So far, only 1 research paper has provided information about classification and its components. Yet, it has been identified that classification plays a crucial role in healthcare data synchronization that helps manage healthcare data with accuracy and efficiency.

Support Vector Machine-based classifier, CNN, and Backpropagation (BP) are the types of classification that have been least included in the literature. Only 1 publication out of 30 journals discussed classification and the term data classifiers. Thus, it can be stated that these terms are the least reliable and capable of improving the research.

The conclusion of this section includes various processes and ways. In other words, sensing healthcare parameters and monitoring data through the cloud environment, data storage through the analysis of available or required storage, and healthcare data communication through classifying data using classifiers. These components are essential for improving elderly healthcare, which the research needs to detail.

8.3 Describing the few publications

In this research, 30 publications included information about sensing, but only 2 articles included a discussion about the cloud. In [31], they proposed cloud-based smart devices that help communicate the healthcare data sensed through the various wearable sensor technologies. In [47], they stated that cloud storage and communication could be used for communicating healthcare data for regular monitoring and tracking of data. Thus, the cloud is used as a tool for healthcare data sensing.

  1. 1.

    Data storage – Storage analysis: All 30 journal papers that are included in the research discussed data storage, but it has been identified that only 1 journal has provided information about storage analysis. Towards that discussed storage analysis using cloud computing to manage extensive healthcare data. Storage analysis is also helpful for effective healthcare data communication [9].

  2. 2.

    Data communication – Classification: The data classification includes a Support Vector Machine-based classifier, CNN, and backpropagation (BP). These terms are included in only one publication, i.e. [14], introduced data classifiers for classifying healthcare data and communicating it with the respective healthcare professionals.

In the research, various kinds of literature were selected that have aimed at data communication, but the classification component was found to be the least. Although there are multiple classifiers for classifying healthcare data, the publications and researchers do not discuss these sub-components.

The fewer occurrences of selected components and sub-components in the discussion section revealed that cloud, storage analysis, and classification have no primary role in apprising monitoring and tracking healthcare parameters for improving elderly healthcare management in the context of the healthcare industry.

9 Conclusion

The research focused on cultivating a model for Health care Monitoring and Tracking through Wearable Sensor Technologies in the context of Elderly Assisted Living. Moreover, it eliminated the issues and limitations of the currently used literature and publications. In addition, the proposed taxonomy (SDD) tested the system in the context of monitoring and tracking accuracy and effectiveness while communicating the sensed data within the healthcare industry. The finding of the system is indicated through the combination of numerous wearable sensors and devices for healthcare parameter monitoring. Further, communication of monitored data through data communication methods is accurate and capable of using feasible tracking and communication of data through a cloud environment. Yet, the system's limitations are imprecise monitoring of healthcare parameters, and healthcare data has not been classified accurately in the research and least reliable. On the other hand, the models have not been tested and verified as of now, leading the research to be aimed at the future direction. The future direction of the research shall aim to test and verify the model in a real-time environment; this can identify the usability and applicability of the system. Moreover, the system should be able to detect numerous tools for categorizing the tracked and monitored vital information.