RELAS: A reliable authentication system for patient monitoring using IoT

Abstract

Owing to ignorance of self health status and lack of regular check-ups, a large number of persons suddenly die in our society. Immediate physical presence of the injured or affected person at a health center for emergency service is also difficult at times. Through advancement in communication and information technologies, Internet of Things (IoT) uses a collection of applications to bring the passive mode healthcare system to the next level for providing real-time qualitative patient care. It incorporates several body sensors to trigger the service to a patient based on his/her current health status. The performances of these tiny sensors are bottlenecked with their limited energy and less computing capability. These constraints present a challenging situation for continuous monitoring of a person. We propose a RELiable Authentication System (RELAS) in which the entire process is divided into four different modes and each mode can have different subtasks and they are capable of executing their tasks either locally or in a server. The process supports both sequential as well as parallel way of execution, which includes periodic collection of physical parameters, environment details of the patient and stores the data in a server for forecasting. RELAS considers all the constraints during the execution period to reduce the battery consumption so as to maintain uninterrupted monitoring of a person. It is observed that the proposed model ensures less consumption of battery power, reduced time complexity and can handle large data size.

Keywords

RELAS IoT NodeMCU PMS real-time

1. Introduction

The growing need for self-regulating healthcare supervision has altered services from infirmary-centric to patient-centric, where a patient can be observed and his/her health information can also be recorded irrespective of location, activity and time. Such type of service provisioning is constrained with bandwidth requirement, data size, consumption of battery and processing speed. In a report [1] of the United Nations it is predicted that by end of 2050, the population of elderly people would be around 22% of the entire world population. It is also expected that 89% of elderly people would be living alone. According to medical survey, 80% of elderly people over the age of 65 might suffer at least one disease. Currently, several healthcare solutions with advanced technologies are available for elderly people [2]. Hence a lot of mechanisms are being proposed by researchers to address the problems for online as well as offline users. However, there is no significant contribution towards extension of time of monitoring the patient by extending the battery life of the sensor with increased data handling capability.

Figure 1.

Model for IoT based healthcare system.

The existing health monitoring approaches are based on selection of the medicine by predicting the disease at an initial stage. Instead of diagnosis and cure most of the approaches are health management oriented [3]. In all cases the physical presence of the patient, personalization-based treatment and management are observed. To overcome this difficulty, our designed model adopts a patient-centric, ubiquitous approach. Service provisioning to a person irrespective of one’s location and time is a challenging task due to mobility of the user among different networks. It is being addressed by providing Body Sensor Networks (BSN) and their integration in the IoT infrastructure. It monitors and supports the person in real-time with the deployment of low-powered, light-weight sensors [1, 4]. The presence of tiny sensors on the body of the person capture the data from the person as well as from the environment periodically which helps to predict any mishap of the concerned person.

Several cases have been reported at hospitals in case of unattended elderly people living isolated life at their residences. Real-time monitoring as well as service provision to those people in our society is a great challenge. In a few cases the devices used to measure the health parameters are not IoT enabled. In this paper an energy aware health monitoring model “RELAS” has been proposed which includes size of data, amount of battery consumption, processing speed and network capacity. The designed model comprises of multiple modules where communication between any two modules follow all-pairs least cost path so as to reduce the battery power consumption and to enhance network life time [5]. In principle, the model ensures efficient processing with respect to accuracy and execution time. Data processing supports both local (microcontroller-based) or global execution (server-based) and it is conditional specific as shown in Fig. 1. The healthcare system consists of a data acquisition system along with a data processing unit to provide services through gateway and different storage devices. Also by including environmental parameters the RELAS provides accurate data to the concerned users by reducing the errors.

2. Related work

“Smart wearable body sensors for patient self-assessment and monitoring” shows the study of the development of Smart Wearable Body Sensors (SWS) and their capacity to reform the framework of medicinal services [6]. It also shows different SWS based systems which are capable of monitoring blood pressure, blood oxygen saturation (SpO2), heart rate, respiratory rate, glucose concentration level in the blood and some neurological parameters. The study concluded that fusing SWS into routine care of patients would enhance relationship between doctor and patient with real-time service as well.

A framework for Smart Hospital System (SHS) has been proposed by Catarinucci et al. [7], where the system can monitor the location of patients as well as capture the biomedical information within the hospital by integrating WSN with RFID. All these technologies are combined together with the help of a network infrastructure built on top up of CoAP (Constrained Application Protocol), 6LoWPAN (IPV6 over low power wireless personal area network) and REST (representational state transfer). They have reduced the power consumption by using zero-power RFID based transmission and ultra-low-power HSN (Hybrid Sensing Network).

The SHS is able to monitor heart rate, patient movement, room temperature and humidity at real-time and delivers to the control room for monitoring. Although the specified model includes many characteristics but it is very costly due to the use of individual RFID devices for tracking individual patient in a healthcare. It is a fixed approach pertaining to a certain area and has less scope for extension to different areas with mobile service.

The Pervasive Patient Health Monitoring (PPHM) system proposed by Ghanavati et al. [8] presents a three-tier PUSH-PULL communication-based model. It preserves both the benefits of IoT and cloud computing. The three layers are collection station, data center and observation station. To measure the accuracy, scalability and energy efficiency of the model, they have considered a patient for congestive heart failure (CHF) testing. The results obtained by them are well accepted. In this model the authors have emphasized the use of cloud server over Wireless Body Area Networks (WBAN). However it missed to include the environmental condition, which is an essential factor affecting the recorded health information of a person.

The article “Wireless sensor networks for personal health monitoring: Issues and an implementation” provides an overall idea about the issues of wireless sensor networks by keeping their focus on Wearable Wireless Body/Personal Area Network (WWBAN) [9]. They have described the general architecture of WWBAN and the key issues related to it such as hardware architecture, software architecture, network time synchronization and energy consumption. The WWBAN based prototype has been designed with custom ECG and motion sensors. The researchers put effort to improve the Quality of Service (QoS), security and reliability of the sensor nodes. Further, it cannot record multiple activities of one sensor. In addition to that, the use of both multiplexer and Wi-Fi against ZigBee improves the wireless range of usability.

Santos et al. [10] proposed an IoT based m-health service architecture using RFID tags [10]. The model implements RFID identification, object name service and Electronic Product Code (EPC) to identify each mobile health related item such as – patient, doctor, medicine, hospital, pharmacy etc. The work also focused on the security context of the software application, access permission of users and control of the devices used. The main objective of this model is to help elderly people by providing a self-managed healthcare. It is only limited to identification of patient but can be improved with real-time health parameters. Also the use of different RFIDs distinguish only different human beings and objects used in a healthcare but cannot assist in the diagnosis of health parameters.

In the “Real-Time Health Monitoring System on Wireless Sensor Network” the authors have designed an IoT based model to monitor the heart rate and oxygen saturation in blood [11]. The model is made up of three sub systems-sender, receiver and data logger. The sender consists of LEDs, PDI-E832 and ZigBee module. The receiver consists of a PIC18F87J10 microcontroller and a ZigBee module. This receiver unit is connected to a computer by using UART. Senders collect the data and send it to the receiver, and the receiver forwards it to the PC once in a second for storage. The model is of affordable cost and provides support through one sensor. As patient monitoring requires gathering of different types of data from variable ranges, hence multiple sensors need to be deployed instead of ZigBee.

Tebje et al. have proposed a low cost sensing system using IoT for monitoring domestic conditions. They have presented integrated and interconnecting network architecture for reliable measurements using sensors and used internet for transmission of data [12]. In their model they have used the longitudinal learning system which provides a self control mechanism for superior operation of things in the observation step. They have combined sensing units with information system for data aggregation to enhance the reliability. It has been observed from the model that the real-time environment data can be transferred to base station through neighboring sensor nodes and it can be utilized for predicting health status of a patient. Data retrieval is costly and can be refined through Wi-Fi technology and a cloud based server.

It has been visualized that IoT devices generate continuous flow of data which can be considered as BigData. The huge amounts of structured and unstructured data, generated by the sensors’ sensing system are transmitted to some cloud server, for further processing. Prabal and Sood have designed a cloud-centric mobile-healthcare framework to monitor and diagnose diseases with security [13]. In their work a UCI dataset has been used to generate student related health status to determine the level of severity of disease. The framework was tested under different classification algorithms and analyzed with respect to its accuracy, sensitivity and measurement. The analyzed data is stored in the cloud and retrieved on-demand, which is not suitable for real-time alert. While in our proposed model after data acquisition it is stored and analyzed in the cloud based server.

In emotion detection based system, anger can be considered as an important parameter due to its variation in values and impact to the lifestyle of an individual. In this context Vivekanand et al. have developed a wearable anger-monitoring system which is used to capture values at different situations by analyzing different activities [14]. It is designed to read different states of anger through a global system of mobile communication. The significance of this model is that it can be wearable throughout the day to determine the mood of a person. However, human behaviour gets affected by environmental parameters. In addition to that we have used Wi-Fi technology instead of GSM, which reduces the cost of the model.

A mobile monitoring system with modular programming has been suggested by Mohammad et al. with adaptable battery power and network availability [15]. They have divided the entire system into different sub processes and implemented the model in parallel environment for faster execution. They have also used a dynamic programming method to obtain an optimal solution. Using this model individual data is being sent to the local sever. Unlike this model we have a used data-fusion method for transmission of the data by maintaining synchronization between sender and receiver.

For real-time heart monitoring, Chao et al. [16] have developed a system which is composed of data acquisition part and transmission part. They have put the idea that acquisition can be done through monitoring all required parameters along with their frequencies. It has been implemented after several interactions with medical experts. It collects several parameters of patient along with location and sends the captured data to the registered users only. Bluetooth controller has been used for data transfer within a restricted area. Along with this they have diagnosed a single parameter for presenting the patient status. Also Al-Aubidy et al. [17] have also proposed a real-time health monitoring mechanism model where different parameters have been captured and are used by the microcontroller to find the status of the patient. The accuracy of collected result is being verified by comparing the result with the result of healthcare device. Applying RELAS on this model its performance as well as accuracy would be improved due to the impact of environmental parameters. The reliable and authenticated status of a patient can be viewed through internet with any web enabled device.

The main aim of our research work is to design and develop a cost effective model to monitor different health parameters of a person and to intimate the person concerned so as to avoid mishap. Periodic monitoring of the heart rate of a person would indicate the random changes in the beats per Minute (BPM). Many researchers have shown the association between the Heart Rate and Body Temperature as $\Delta$ HR/ ${}^{\circ}$ C of 14.7 BPM/ ${}^{\circ}$ C. Also it has been accepted the “V” shaped relationship between room temperature and the heart rate. Also to find accurate parameter of a patient external temperature needs to be considered to get physiological status and heart rate of a patient. So we have included body temperature, room temperature and humidity in the monitoring list, as it will help any specialist understand the actual scenario of the patient. In our work, we have successfully built the model “RELAS” to monitor the data and to communicate the authenticated data to the concerned persons within a desired time. Finally, to show the potential of our model RELAS its performance has been compared with existing models as shown in Table 1.

Table 1
A comparative analysis of different models

Model	Multiple sensor attached	Remote healthcare solution	Includes environment conditions	Portable model	Wi-Fi supported
Smart hospital system (SHS)
Pervasive patient health monitoring (PPHM)
Wireless sensor networks for personal health monitoring: Issues and an implementation
Internet of things and smart objects for m-health monitoring and control
Real-time health monitoring system on wireless sensor network
Towards the implementation of IoT for environmental condition monitoring in homes
Cloud-centric IoT based disease diagnosis healthcare framework
Wearable anger-monitoring system
Resource-aware mobile-based health monitoring
The IoT-based heart disease monitoring system for pervasive healthcare service
Real-time healthcare monitoring system using wireless sensor network
Proposed model (RELAS)

3. RELAS: The proposed model

Our proposed model “RELAS” operates in four different modes. The four modes are (i) Instantaneous unremitting communication mode, (ii) Unremitting communication in exceptional phase mode, (iii) Event generated communication mode and (iv) User’s request mode. Data from different sensors are collected by a data acquisition system, processed, and then forwarded through gateway to the cloud server for storage and further processing. We have attached wearable body sensors along with intelligent microcontroller for data acquisition, and the data is stored in cloud-based storage to provide access to the authenticate users irrespective for their time and location in the globe. The stored data can be immediately forwarded to the registered users of healthcare system to analyze and to detect any variation in person’s health parameter values. Figure 2 represents the proposed model whose characteristics are described through all four different modes.

Figure 2.

Design of real-time patient monitoring system (RELAS).

The four different modes of operation of our model is being outlined as below.

3.1 Mode-1: Instantaneous unremitting communication mode

Through this mode the entire data will be delivered from source sensors to the server for storing as well as to the remote device for displaying the authenticated data. Following this mechanism the sensors will gather the original data from the patient body as well as environment and transmit through an intelligent interface to the server through various connectivity. This mode is the essential for those patients who belong to a high risk health status like acute stroke or coma patients and need continuous monitoring along with healthcare diagnosis using different protocols. So it provides the required features through sensors and microcontroller to the end user through forecasting to get a quick response. The processing cost can be evaluated by considering a few factors associated with this mode.

Microcontroller device processing speed (bytes/ second) $=P_{\rm m}$ . Server processing speed (bytes/second) $=P_{\rm s}$ . Mobile device processing speed (bytes/second) $=P_{\rm mb}$ . Consider $S_{1},S_{2},\ldots,S_{k}$ are the steps for completion of communication of all patients. Distance covered during processing n bytes of data at step $S_{i}=P_{i(^{n})}$ . Time to process (seconds) n bytes of data at step $S_{i}$ in the server is ${t_{i}}^{s}(n)=\frac{P_{i(n)}}{P_{\rm s}}$ . Time to process (seconds) n bytes of data in microcontroller is ${t_{i}}^{\rm m}(n)=\frac{P_{i(n)}}{P_{\rm m}}$ . Time to process (seconds) n bytes of data in mobile device is ${t_{i}}^{\rm mb}(n)=\frac{P_{i(n)}}{P_{\rm mb}}$ .

But in reality this mode transmits maximum data and stores it at the server side, which makes the volume large. For this, the requirement of good network quality with respect to bandwidth comes with increase in data size and also the battery consumption will be maximum that makes the network lifetime minimum. Still, as shown in Fig. 2 this mode helps in high level monitoring with instant attention from healthcare system; during service provision the number of patients will be limited due to minimization of network lifetime, so the next mode can be a solution to this mentioned issue.

3.2 Mode-2: Unremitting communication in exceptional phase mode

It has been suggested through this mode that continuous sensing of data will be done through sensors and can be stored in server; only the exceptional data will be considered for forecasting. From the medical surveys it has been deduced that heart attacks may take place at multiple special times, like 1 or 2 hours after getting up, and 3pm to 4pm due to heavy stress [18]. For exceptional situations like this the monitored data can be forwarded in real-time so that a quick response can be received from the concerned doctor during this time period. For completion of operations, sequential execution passing through all devices needs to be done.

$E_{x}$ denotes the execution status storage as that parameter is used for two different purposes i.e. whether step $S_{i}$ executes on microcontroller or $S_{i}$ executes on server.

$\displaystyle E_{x}=\left\{\begin{array}[]{ll}0,&\text{if }S_{i}\text{ % executes on microcontroller}\\ 1,&\text{if }S_{i}\text{ executes on server}\end{array}\right\}$

So the execution time of individual sections like sensing and transmitting by sensors, time consumed for execution by microcontroller, server and mobile separately are represented respectively as

Execution time for one step $t_{i}(n_{i})=E_{x}t_{x}(n_{i})+(\overline{E_{x}}){t_{i}}^{\rm m}(n_{i})+(E_{x}% ){t_{i}}^{s}(n_{i})+{t_{i}}^{\rm mb}(n_{i})$

Total execution time

$\displaystyle T(n)=\sum^{k}_{i=1}{t_{i}(n_{i})}=\sum^{k}_{i=1}(E_{x}t_{x}(n_{i% })+(\overline{E_{x}}){t_{i}}^{\rm m}(n_{i})$ $\displaystyle\qquad\quad∼{}+(E_{x}){t_{i}}^{s}(n_{i})+{t_{i}}^{\rm mb}(n_{i}))$ $\displaystyle\qquad∼{}=T^{\rm m}(n)+T^{s}(n)+T^{\rm mb}(n)$ $\displaystyle T^{\rm mb}(n)=T(\text{Constant value})$ $\displaystyle T^{s}(n)=\sum^{k}_{i=1}{(E_{x}){t_{i}}^{s}(n_{i})}$

As transmission time is common for microcontroller and server, so

$\displaystyle T^{\rm m}(n)=T(n)-\ (T^{s}(n)+T^{\rm mb}(n))$

3.3 Mode-3: Event generated communication mode

This mode specifically can be executed in parallel for offline as well as online users. If the patient data goes beyond the threshold value and any event occurs against the patient, then notification will be given to all offline as well as online authenticated users. To achieve this mechanism in our model for offline users, we have used a buzzer to provide the signal of event and through mobile application in parallel the notification can be sent to all related authenticated users for further assistance. So either the communication will be done at the server end or mobile end depending on the phase. This mode minimizes the data to be sent to the healthcare system, thus improves the system and lightens the pressure at server side. This mode also suits all healthcare people very well as they cannot monitor the data for the entire time but can assist once they get a notification as emergency. As a body temperature sensor will provide a value of patient including surrounding temperature value, so using the environment temperature sensor the actual body temperature value of patient can be calculated using the model. So, if in this phase any calculated value crosses a threshold value then the appropriate personnel is notified instantly.

3.4 Mode 4: User’s request mode

Figure 3.

Circuit diagram of the RELAS model.

This mode is typically used for fulfillment of user’s requirement on demand. As the data are stored in server for future use, and on request status is shown to patient and other authenticated users. This mode is considered as the lowest level of health monitoring mode mostly because it fits patients with very low health risk. Basically the doctor and other authenticated users related to patient need to check the status of diagnosis, so when necessitated data from the server will be forecasted to the users.

4. Hardware configuration of RELAS model

Figure 3 shows an eLua based open source IoT development board NodeMCU being used in our real-time model within which ESP8266 has been integrated to enable access point mode and stationary mode operation. It has 10 GPIO pins D0 D10, one ADC pin and several I2C and SPI communicators. All the components have several pins with different functionalities and have their own pin configuration. As all pins are not in action phase, the required are connected for the complete description of proposed model through pin configuration as shown in Fig. 2. Both LM35 and Pulse sensor provides analog signals as output where NodeMCU-12e is built in which only one ADC pin is used both for input and output operations. So, CD4051 is attached as an 8:1 multiplexer to select the outputs of both Pulse sensor and LM35 in a serial manner. It has 3 selected pins and based on the combination of these select pins the data is collected periodically from the respected module and forwarded to the destination module. The single output pin of CD4051 is connected to the ADC of NodeMCU and is used to forward all captured relevant data through the chip select line, according to which either the output signal of Pulse sensor or the output signal of LM35 is delivered by CD4051 to achieve synchronization. A LCD of size 16 cm $\times$ 2 cm with 16 pin device is used to monitor the patient details locally. Due to the limited number of pins of NodeMCU, we have used an I2C adapter to interface the LCD with the NodeMCU. The LCD has 16 columns and 2 rows where a 6 pin push button on-off switch is used as the mode selector of LCD display. Both the modules DHT11 and DS3231 are connected to the NodeMCU via corresponding GPIO pins, for digital output.

Proposed algorithm[1]

$P_{0}$ : Pulse Sensor Pin number, $P_{1}$ : Temperature Sensor Pin Number, $P_{2}$ : DTH11 Sensor Pin Number, state: State of Switch, $V_{\rm in}$ : Supply Voltage in millivolt, $P$ : Phase Difference $\theta_{\rm lb}$ , $\theta_{\rm ub}$ : Lower and Upper Threshold value for pulse sensor (0 and 1024 respectively)

Beats Per Minute (BPM), Body Temperature, Room Temperature, Humidity, Current Date & Time

i = 1 to 60

time = getTime (DS3231);

changeMUX ();

FirstBeatStatus $=$ false;

signal $=$ getdata ( ${P}_{0}$ );

(signal $<$ $\theta_{\rm lb}$ or signal $>$ $\theta_{\rm ub}$ )

display “Status Normal”

(FirstBeatStatus $==$ false)

FirstBeatStatus $=$ true;

FirstBeat $=$ getTime (signal);

SecondBeat $=$ getTime (signal);

FirstBeatStatus = false;

RELAS[] $=$ SecondBeat $-$ FirstBeat

BPM[i] $=$ 60000/avg (RELAS[]);

changeMUX ();

TempData $=$ getValue ( ${P}_{1}$ );

$V_{\text{out}}=$ TempData * $(P/V_{\rm in}$ );

cTemp $=$ $V_{\text{out}}$ /10;

fTemp[i] $=$ (cTemp $*$ 9)/5 $+$ 32;

rTemp[i] $=$ getValue ( ${P}_{2}$ );

Hum[i] $=$ getValue ( ${P}_{2}$ );

FinalTemp[i] $=$ Temp[i] $-$ rTemp[i]

FinalBPM $=$ avg (BPM[])

FinalTemp $=$ avg (FinalTemp[])

FinalrTemp $=$ avg (FinalrTEmp[])

FinalHumidity $=$ avg (FinalHum[])

(FinalBPM && FinalTemp && FinalrTemp && FinalHumidity)

(state $==$ ON)

Operate Patient Details

Operate Environment Details

((FinalBPM $\|$ FinalTemp) $\geqslant$ Threshold)

Buzzer.on ()

Return

5. Cost analysis of the proposed model

The working time for four different modes has been considered to find the total cost of model. The idea from this cost analysis method has been used in the algorithm for finding a better solution to make the model energy aware, as it includes different size of data and its corresponding battery consumption to enhance the network lifetime. Also the algorithm here is specifying the step by step detail procedure for collecting and calculating the desired results for corresponding patients locally and remotely.

It is shown from different modes that the data forwarding and processing cost at the existing step are not dependent on the calculation option selected at the preceding step. The specified preliminary input data size is of $n$ , from which we can differentiate ‘ $n_{i}$ ’ for each step. Also for a request each step is independent of selecting the modes. In our model we have made it static during defining of modes to minimize the cost of battery consumption. The total cost can be calculated by considering the forwarding of previous step as,

$\displaystyle C[i,l,b]=\text{Cost of solution for the mode `l' of step}\text{$% {S}_{i}$ using total battery consumption `b'}=C[i-1,j,b-{d}_{i-1}[j,l]+{e}_{i}% [l]]+{W}_{i-1}[j,l]+{t}_{i}[l]$

where, ${W}_{i-1}[j,l]=$ Transmission cost from mode $j$ of ${S}_{i-1}$ to mode $l$ of ${S}_{i}$ , ${d}_{i-1}[j,l]=$ Battery consumption for transmission from mode $j$ of ${S}_{i-1}$ to mode $l$ of ${S}_{i}$ , ${t}_{i}[l]=$ Execution cost at step ${S}_{i}$ for mode $l$ , ${e}_{i}[l]=$ Battery consumption for execution cost at step ${S}_{i}$ for mode $l$ .

6. Analysis of experimental results

The real-time patient monitoring system (RELAS) has been run through different phases as shown in Fig. 1 from which data has been collected and different services have been given to registered clients after analysis. Various parameters were considered as arguments for representation of resultant graphs by following the architecture displayed in Fig. 3. The patient related information like heart rate, body temperature, environmental conditions as well as current date and time have been collected using medical sensors using local processing unit and this data is forwarded to the RELAS server where different operations are performed related to patient. Finally the patient diagnosis result (PDR) is recorded for providing the status of severity of a patient to all clients.

Figure 4.

Execution time vs input size.

For all different challenging mechanisms, the execution time enlarged sequentially with the given data size. If we will consider the case of execution only in mobile, then for any input data execution time is the maximum as shown in Fig. 4. Considering the case for contribution of 100 KB data, running time of mobile is calculated as 1475 seconds, Similarly of data size 1024 KB, almost 10 times of original time, i.e. 147018 sec is consumed. The parallel mechanism consumes the significantly lesser execution time i.e. 971 sec for 100 KB and subsequently for 1024 KB input its 9699 sec. The subsequent lesser execution time is provided by using server approach then lower is the optimal mechanism. For input of 100 KB, server consumes 195 sec time, where optimal solution consumes 1172 sec. For input of data size 1024 KB, the server execution time is 1735 sec and optimal solution with respect to time is 1222 sec. The cause for maximum execution time of mobile is observable as every execution is carried out using mobile only mechanism and this execution time is much more than the server only approach. The parallel mechanism provides lesser execution time compared to mobile approach as execution are carried out in server and mobile in parallel. The mechanism of server only also shows lesser execution time as every execution has been completed using server only, thus is quicker than parallel execution as little part in mobile and few are at server is completed. Although the optimal solution model consumes lower time, but in RELAS real-time model, it considers each step of execution including maximum work at microcontroller level and then distributes the same among other parts for making execution faster, so it consumes 112 sec for 10 KB data and subsequently for 1024 KB it takes 1176 sec. It has been calculated from the real-time generated result of patient and is being compared with other models in Fig. 4. As authors have designed the real-time model to collect data continuously, so supportive calculation for the same is also mentioned. But during no transmission the model will be off and will not consume the same battery like other models because the input data to the model will be of 0 sizes.

Figure 5.

Battery consumption vs. input size.

As the execution time is indirectly proportional to the battery life as well as directly proportional to the energy consumption, so feasibility of RELAS model is being verified by comparing with other models through battery consumption and analyzed results are shown in Fig. 5. The calculation of battery consumption is calculated as like, let battery utilization per sec of CPU procedure $=$ d units. Consider $t_{0}$ as the execution time for mobile only approach without including the developed algorithm and on RELAS the algorithm is applied on the equivalent data. So, energy utilization by mobile $=$ dt ${}_{0}$ , and by RELAS $=$ dt ${}_{\rm pms}$ . As a result, decrease (%) in energy utilization by RELAS as compared to Mobile $=$ 100 (dt ${}_{0}-$ dt ${}_{\rm pms}$ )/(dt ${}_{0}$ ) $=$ 100 (1 $-$ (dt ${}_{\rm pms}$ /dt ${}_{0}$ )). Referring to Fig. 4 the time to execute data of 1 MB size by mobile mechanism is 14718 sec [15] and time taken by RELAS $=$ 1176 sec. Therefore, % of decrease in energy utilization for RELAS in place of mobile $=$ 100 (1 $-$ (1176/14718)) $=$ 92.01%. In additional meaning, for 1 MB data size mobile only mechanism consumes 100% battery, Optimal consumes about 9% [15] of battery and RELAS uses only 8%. Also considering different size of data, RELAS provides better results with less battery consumption than other methods.

The model has been tested under three different stages of a patient for real-time monitoring and able to capture the authenticated data for all four different parameters of a patient. To validate the monitoring capability of the developed RELAS model, we have used it on the patient’s body and captured the parameters as signals transmitted under different daily activities and found corrected data as compared to the manual model. The patient has been observed during rest at home, exercise and playing sports at afternoon, morning and evening respectively. Figure 6 illustrates the result of the heart rate, body temperature along with environment conditions of one user at different stages. This result can be shown to all registered users as per request at remote location in real-time using the designed RELAS model. So as per Fig. 6 the proposed model can be used by anyone at any place to get the connected users current health status.

Figure 6.

Monitoring of multiple parameters of a single patient during different activities.

Figure 7.

Real-time monitoring of different parameters of multiple patients.

RELAS model also has been tested for multiple users to find the authenticity which is represented in Fig. 7. It shows the frequencies of occurrences of different parameters for five different users at different time intervals. Out of all users it was found that patient-4 and patient-5 have reached a high pulse rate along with excess body temperature and for which automatic notification has been intimated. The potential of the system to classify urgent circumstances according to the previously set threshold values has been examined by using captured data from the model and instant auto generated information is also shared with all authenticated users associated with the patient. All generated data are stored in a cloud server for real-time monitoring and diagnosis by user itself and healthcare users respectively so that any kind of problem can be avoided before time. Using this model the death rate of patients can also be decreased.

The data points are statistically independent as four different parameters are collected to get the status of patients. The algorithm has been designed in such a way that, with the help of CD4051 we will get all four parameters one after another at a single point of time and the model will transmit all four parameters to the server. After successful delivering of the data, the device will go for next loop to get the data of the next time interval. As shown in Fig. 3 the RELAS model measures the parameters at different time intervals of the patient in various situations which again undergoes different stages of processing and displayed as an analysis of results through Figs 6 and 7. The same model is used to collect data related to multiple parameters of a single patient as well as from several patients to make it cost effective.

7. Conclusion

The progress in IoT mechanisms and the advancement in network connectivity protocols for devices enable a real-time health monitoring system to be cost-effective and attractive. The continuous monitoring system generates huge amount of data and also consumes battery power of the sensors to a large extent. Our proposed model RELAS exhibits a reliable, authenticated and cost-effective framework for real-time patient monitoring with optimal accumulated data with reduced energy consumption. It integrates smart properties using different modes to address the problems of managing large continuous traffic, battery usage and also provides real-time solution for offline as well as online users. It has been shown through various experimental results that RELAS uses different computation modes, accepts large input, consumes less battery and exhibits less time complexity. RELAS can effectively collect information from multiple patients with accurate evaluation, and ensures authenticity during monitoring the health status in real-time. With the integration of nanosensors into our model the size, scalability and performance can be improved with industrial viability. This model would also be useful for patients at remote places as well as places with restricted entry.

References

World Population Ageing, 2013, United Nations, New York, NY, USA, (2013), 8–10.

Weinstein

, RFID: A technical overview and its application to the enterprise, IT Prof 7(3) (2005), 27–33.

Istepanian

R.S.

Philip

N.Y.

and Sungoor

, The potential of Internet of m-health Things “m-IoT” for non-invasive glucose level sensing, in: Engineering in Medicine and Biology Society, EMBC, 2011 Annual International Conference of the IEEE (5264-5266), 2011.

Yang

Xie

Mäntysalo

Zhou

Pang

Da Xu

and Zheng

L.R.

, A health-IoT platform based on the integration of intelligent packaging, unobtrusive bio-sensor, and intelligent medicine box, IEEE Transactions on Industrial Informatics 10(4) (2014), 2180–2191.

Prosanta

and Hwang

, BSN-Care: A secure IoT-based modern healthcare system using body sensor network, IEEE Sensors Journal 16(5) (2016), 1368–1376.

Appelboom

Camacho

Abraham

M.E.

Bruce

S.S.

Dumont

E.L.

Zacharia

B.E.

and Connolly

E.S.

, Smart wearable body sensors for patient self-assessment and monitoring, Archives of Public Health 72(1) (2014), 28.

Catarinucci

De Donno

Mainetti

Palano

Patrono

Stefanizzi

and Tarricone

, An IoT – aware architecture for smart healthcare systems, IEEE Internet Things J 2(6) (2015), 515–526.

Ghanavati

Abawajy

J.H.

Izadi

and Alelaiwi

A.A.

, Cloud-assisted IoT-based health status monitoring framework, Cluster Computing 20(2) (2017), 1843–1853.

Aleksandar

Otto

and Jovanov

, Wireless sensor networks for personal health monitoring: Issues and an implementation, Computer Communications 29(13-14) (2006), 2521–2533.

10.

Santos

Macedo

Costa

and Nicolau

M.J.

, Internet of things and smart objects for M-health monitoring and control, Procedia Technology 16 (2014), 1351–1360.

11.

Deepak

Kumar

and Gupta

, Real-time health monitoring system on wireless sensor network, International Journal of Advance Innovations, Thoughts & Ideas 1 (2012), 37–43.

12.

Tebje

K.S.D.

Suryadevara

N.K.

and Mukhopadhyay

S.C.

, Towards the implementation of IoT for environmental condition monitoring in homes, IEEE Sensors Journal 13(10) (2013), 3846–3853.

13.

Prabal

and Sood

S.K.

, Cloud-centric IoT based disease diagnosis healthcare framework, Journal of Parallel and Distributed Computing 116 (2018), 27–38.

14.

Vivekanand

Prakash

and Sagar

, Wearable anger-monitoring system, ICT Express (2017).

15.

Mohammad

M.M.

Serhani

M.A.

and Navaz

A.N.

, Resource-aware mobile-based health monitoring, IEEE J Biomedical and Health Informatics 21(2) (2017), 349–360.

16.

Chao

X.P.

and Zhang

L.L.

, The IoT-based heart disease monitoring system for pervasive healthcare service, Procedia Computer Science 112 (2017), 2328–2334.

17.

Al-Aubidy

K.M.

Derbas

A.M.

and Al-Mutairi

A.W.

, Real-time healthcare monitoring system using wireless sensor network, International Journal of Digital Signals and Smart Systems 1(1) (2017), 26–42.

18.

Plotkin

Vaibavi

S.R.

Rufaihah

A.J.

Nithya

Wang

Shachaf

and Seliktar

, The effect of matrix stiffness of injectable hydrogels on the preservation of cardiac function after a heart attack, Biomaterials 35(5) (2014), 1429–1438.