Artificial Intelligence Project



Intelligent Health Maintenance System

Predictive Health monitoring and Maintenance based on Cloud Computing and IOT Architecture using Machine Learning














Chapter 1: Introduction. 6

Problem Definition: 6

Background: 8

Strategy: 9

Research Objectives: 10

Research Limitations: 11

Chapter 2: Literature Review.. 13

Chapter 3: Research Methodology. 18

Chapter 4: End-To-End Architecture For Enabling Artificial Intelligence Through Stream Analytics  20

Stream Analytics: 23

Machine Learning: 24

The Microsoft Azure Machine Learning Studio: 26

Training And Scoring Models: 28

Heart Disease Prediction: 29

Chapter 5: System Requirements. 34

Functional Requirements: 37

Non-Functional Requirements: 42

Project Issues: 44

Sysml Requirements Diagrams: 47

Diagram 1: Operating Environment: Indoor And Outdoor 47

Diagram 2: Type Of Alarm Depending On The Movement: 48

Use Cases: 49

Business Use Cases: 49

Product Use Cases: 55

Adapt Case: Enhancing Data Analysis And Decision Making For Fall Detection Using Machine Learning Algorithms. 60

Chapter 6: Conclusion And Recommendation. 64

References: 67

Appendices: 80

Appendix 1: 80

Appendix 2: 84

Appendix 3: 85

Ploting Correlation Matrix. 86

Binary Classifier Creation. 86

Model Accuracy Evaluation. 86




Chapter 1: Introduction

Problem Definition:

Ever since the advent of smart devices which can be accompanied with humans in mobile forms and be connected to the internet, medical industry has been dreaming of a comprehensive integrated system through which health can be monitored and diseases can be prevented using the data from health monitoring devices. Human health has been at the mercy of doctors and health practitioners, due to which errors are probable in diagnosis of any medical condition. The capability of cloud computing with its ability to handle big data and web of smart devices can give a way for predictive diagnosis which can possibly eliminate human error in prescribing medicines for diseases and as a preventive mechanism for future health threats.

The modern healthcare facilities generate tremendous amounts of data which is handled through conventional in house e-healthcare systems. However, such systems does not have the capability to integrate and process data streams in the form of big data from many different dimensions and sensory devices distributed in hospital all the way to the patient’s home. Healthcare facilities use the traditional relational database management system (RDBMS), which is limited in its processing power and scope to make sense of big data which requires processing in real time such that appropriate steps can be taken before the problem jumps out of reach. In addition to the stated limitation, as mentioned before the new healthcare system is shifting from hospital-centric or career-centric to patient-centric, where patients are to be monitored periodically throughout the day and they themselves along with medical professionals can be empowered to make best decision by taking in view the processed information.  The volume and speed at which there is a requirement of organising the sensor data to make accurate prediction is unprecedented for traditional RDBMS. The big clusters coming in data streams from patient’s home need to be streamed and analysed in real time with seamless connectivity.

With an increase in the use of Internet-of-Things, the data generated from every patient comes in clusters of big data, which is extremely useful for making health situation and cure related prediction if it can be processed and analysed timely. It is very difficult to make sense of data streams in real-time through manual analysis of health care professionals; hence, highly complex computing algorithms are the only possible methods to abstract and make sense of clusters using historical data and enable machine learning of the system.

The system shall cover all the required technicalities which are appropriate to design a cloud Computing based intelligent health monitoring and prevention system for the management and usage of health care data that will use predictive analytics to assess patient’s health condition and suggest appropriate solutions to both patients and health professionals from remote locations along with suggesting ways to prevent a possible health condition in the future. In recent past many systems have been proposed but they all have been designed without the capability of completely eliminating human involvement in drug prescription and use web of devices to monitor health and diagnose according to trends through machine learning at the same time. In the above stated lines, the system is intended to provide web based user friendly platform to patients in order to empower them to monitor their health condition or symptoms, have a deeper insight into the symptoms through the use of recommendations and analytics enabled by machine learning algorithms; receive health professional validated recommendations and keep track of their health on a visual website.




In the coming decades, the ageing of population and the revolution of Internet-of-Things (IoT) are among the megatrends that will reshape the world dramatically (National Information Council 2008). According to the department of Economic and Social Affairs of United Nations (2009), the rise of the persons above 60 have been steady during the recent years, which has incremented to 115 in 2009 from 9% in 1950; which according to the recent trends can go as high as 25% in 2050. The aging population rings bells for the already struggling healthcare systems around the world, which lack the robustness to handle huge influx of patients in the coming years. According to National Bureau of statistics (2011) developing countries such as China are even more disposed to this problem. In addition to the rise in population of elderly the overall population in the developing countries is increasing which is partly because of a higher mortality rate; combined with the increasing vulnerability of human health caused by pollution and other artificial ways of everyday lifestyle, every individual is much more disposed to viruses and disease. The best way to handle this rising problem is to introduce ways that have the capability to encourage preventive measures such that any disease can be prevented. At the same time the healthcare system shall have the capability to reach maximum number of individual and monitor their health to enable a transformed ecosystem for healthcare powered by the modern technologies. The future health care services shall be transformed towards patient centricity than the previous criteria of career centricity. Cloud computing provides the required infrastructure which can make the healthcare system completely patient-centric and transparent with a greater degree of reliability and robustness. The machine learning capabilities of the cloud infrastructure and the integration of Big Data analytics can allow the system to predict trends in the patient’s health and prevent any disease by suggesting the appropriate measures based on reliable accumulated data. Such system can be realized through te interjunction of Cloud computing, Data Science for acquiring, processing and managing data for stakeholders in Healthcare.

The current healthcare system is headed towards home-centricity, targeted for 2030 through hospital centric model which is currently dominant. According to Koop et al. (2008,29): Technology will transform our medical system by bringing technology directly into the home. The enabling technology is in two directions: Cloud computing and internet of things. IoT enables the cyber-physical system controlled by cloud infrastructure which has the capability to remotely process information and deliver it to the smartphones of the patients and healthcare professionals. The system has the capability to integrate devices and applications for communication, sensing, information and networking management, identification, through any media and device, anywhere anytime; morevover, the system can be made fully intelligent through the integration of predictive analytics and for the very first time in the history a system much more capable than human mind dedicated to health monitoring and maintenance can be realized and implemented (European Commission Information Society 2008, 1009).



The study is focused on utilizing trans-disciplinary approach to probe into the latest computational infrastructure of Cloud and IoT devices to enable Intelligent Health Maintenance System (IHMS), which can monitor current health condition and potentially prevent health deterioration and disease through use of predictive analytics to prescribe preventive medicine based on big data clusters acquired from web of monitoring devices. The system is aimed at catering the in-house health monitoring of the elderly along with providing a health maintenance system for health conscious individuals by acquiring data such as Blood Pressure, Heart Beat Rhythms, Blood Sugar levels, Body Temperature, Physical Activity, indoor air pollution etc. The trends in data acquired for each individual shall be processed through cloud based predictive analytics algorithms using Software as a Service (SaaS) such as Apache Spark and IBM Watson Service. The user generated data shall be compared with cleansed data in clusters through using historical and present trends (predictive analytics) and the trends in the data shall be matched to predict a health condition. The system shall be artificially intelligent given that it will learn with each data sample and will store greater clusters for matching.

The research study is divided into two major stages: first stage aims at collecting data for constructing framework of the system using literature, online courses (MOOCS), and publications of industry leader. The aim here is to probe the technological advancements in multiple computing and health care domains which can make the system truly intelligent through machine learning. In the second stage, the collected secondary and primary data is used to define requirements of the system along with construction of complete end-to-end architecture of the system. The business case for the system is also constructed. The system archtecture modules of the system include: IoT Hubs, Events Hubs, Stream Analytics, Machine learning, Storage and Networking, Web APIs and Business Community Strategy. Special consideration shall be given for the solution of the following issues: Selection predictive analytics platform for accuracy and efficiency; Data Handling issues; Network system security; Ethical and privacy issues.

Research Objectives:

  • To understand the application of machine learning and predictive analytics in cloud based systems.
  • To investigate the needs and requirements of a comprehensive Intelligent Health Monitoring System (IHMMS) using Secondary (content analysis) data and exploit platforms to enable machine learning.
  • To define the system architecture of IHMS through the use of requirements defined using the collected information.
  • To Investigate and define the solutions to technical problems in acquiring and processing data in real time and using predictive analytics through Cloud computing.
  • To identify and provide solutions of possible challenges and obstacles in the implementation of IHMS on a larger scale for consumer market.



Research Limitations:

This research will be completed using primary and secondary data. This means that there will be primary data collection for the purposes of this study from online survey. However, the use of primary data does not mean that resources are not required. The study will require one to use and internet connection and subscribe or buy some publications over the internet. While most of the publications being sought are likely to be offered free of charge, some may require subscriptions to get access to their content.

The study also aims to utilize data from online courses which are mostly free but researcher will have to purchase few resources for the course work. The study also aims at developing a prototype of the IHMS, which requires use of cloud services as Platform-as-a-service (PaaS); hence, the researcher will have to subscribe to cloud services which might be very costly.

Due to the limitations pertaining to the hardware implementation and collection of appropriate data for predictive analytics the system cannot be fully tested with real time data acquisition.

The study is aimed at exploiting the most recent trends in the technology and healthcare, however, such an approach has many constraints given that much of the research and development in the field is not openly published and is constantly being updated and new standards are introduced.







Chapter 2: Literature Review

Yu et al. (2016) proposed an intelligent system based on cloud computing software as a service approach for handling big data in health care. The focus is on building an efficient platform based on web services in order to simplify the procedures of health monitoring and assessment. The system realized uses big-data analytical approaches to provide reliable recommendations through the comparison of cleansed data with the generated by the user of the system. The repository of cleansed data clusters are acquired by the researchers from the IBM Watson and Apache Spark services. The system provides web based user friendly platform to patients in order to empower them to monitor their health condition or symptoms, have a deeper insight into the symptoms through the use of recommendations and analytics enabled by machine learning algorithms; receive health professional validated recommendations and keep track of their health on a visual website. Yu et al. (2016) provided a one-click solution to all the patients which can potentially solve the problem of inefficient diagnosis and delayed medical attention.

Yamamoto et al. (2014) defined a system through the use of cloud computing to manage big data for connected smart cities. The researchers demonstrated ways to retrieve data from “Scallop4SC”-a smart city platform-of the connected smart houses. The researchers utilized HBase for variety of log data and Hadoop to big-data logs. The study concludes that different data processing and mining methodologies are suitable for different sizes of data acquired.

Jiang et al. (2014) constructed a system to cater the transfer of data for Healthcare big-data systems. They used wearable devices to acquire the data and constructed an intelligent Information Forwarder. They focused on constructing the system for the healthcare of elderly. They provided platforms which allow for centralized analysis of distributed monitoring.

Li et al. (2014) defined a system for sensory data acquired through health monitoring devices and proposed a community based wiki-health as a big data platform. The proposed platform defies a new ontology for data information management which includes ways of analyzing, retrieving, tagging, storing, searching and comparing data acquired from health sensors. The system allows individual users to access health-related knowledge discovery through internet. A similar system was proposed by Kang et al. (2016) which is constructed to cater acquired sensory data of supply chain from Electronic Product Code Information Services (EPCIS) using MOngoDB.

Mohammed et al. (2014) focused on building a mobile application for android platform for healthcare domain, which aims at acquiring the sensory information from connected internet of things and processing the data using cloud computing. The name of their application is “ECG Android App” with an aim to provide users a platform to visualize their current health condition through their Electro Cardiogram (ECG) waves in real time in comparison with historical data. A specific medical cloud is also integrated for the uploading of the user’s data anonymously in order to build repository of cleansed samples. The user data is also uploaded to the personal centralized cloud of the user which can be accessed by medical professionals after permission of the user in order to assess and diagnose their health condition. The researcher presented a comprehensive infrastructure for healthcare industry based on integrating many different technologies: signal processing, communications protocols, IOIO microcontrollers, efficient and secure systems for data sharing, signal processing, central cloud and data base management system. The focus of their study is to construct a comprehensive system architecture and design which can enable the integration of cloud and IoT for healthcare industry. The figure below demonstrates the layered architecture of the proposed patient monitoring system by Mohammed et al. (2014).

Figure 1: Patient monitoring system system architecture including all layers (Mohammed, Thakral, Ocneanu, Jones, Lung, & Adler, 2014)

Song et al. (2012) used android smartphone application and hardware to construct a solution for body monitoring system. They focused to solve the problem of unreliability and limited scope of previously available body monitoring system by utilizing the smartphone technology. The information related to the user’s acquired data can be accessed through the smartphone screen. The researchers integrated multiple technologies in the system design: electrocardiogram acquisition sensor, brainwave capture sensor, temperature detection, smartphone application, and network transmission. They further tested the efficacy of the proposed system in real world cases.

Lee et al. proposed the “Concinnity” system which acquires data from multiple sensors and uses a repository along with workflow system to process and interpret the information. They proposed a cloud platform-as-a-service to manage the sensory data and demonstrated the integrated architecture for the system. They integrated the data input portal where the users of the system can contribute their data to the big data repository. The data that can be uploaded includes lifestyle information, sensory data, well-being etc.

The Health-IoT technologies and services are promising to address the challenges faced by the health care sector. However, literature on the subject shows limited work on the integration of the available technologies to mechanize an automated holistic health care and prevention system for general consumers.

The literature review is providing deep insights into the technological requirements of designing a system that can be truly intelligent and incorporate machine learning. The focus is on the cloud computing elements enabling machine learning capabilities of the system; hence multiple domains have been analyzed, such as predictive analytics, machine learning, big data analytics etc.; using different research instruments. The primary research instrument used is the analysis of articles in peer reviewed journals.

The literature on the subject area shows that there is a need for an intelligent health monitoring and maintenance system which can acquire multiple health related factors through mobile devices and use predictive analytics to suggest preventive medicine along with suggestions for the patient’s decision regarding their healthcare. However, the literature lacks a complete cloud based solution, which can integrate live streams of data acquired through the web of IoT and use Cloud SaaS to for stream analytics and machine learning. Such a system can have unlimited processing power and an ecosystem of patients can allow the system to be much more intuitive than any other system presented in the literature given that this system shall have the capability of expansion by using greater number of measurements to derive a conclusion.


Chapter 3: Research Methodology


The research study utilizes content analysis and desk based research strategy with an inductive approach by first distilling the appropriate work in the literature and then constructing thresholds to assess the efficacy of the a Cloud Based HealthCare system.

The dependence on literature as major form of data extraction has its pros and cons, when using the content analysis technique; such challenges are identified by Kalenga (2015), which has been considered throughout the study as benchmark to objectify the information. In order to limit the possible sources of anomalies the study inducts triangulation of key findings through primary data collection from online survey (Kratina et al., 2015).

The aim of the method used is to precipitate qualitative information under a theme followed for the analysis. The quality of the published reports is also assessed based on the standards followed in the industry. Bryman and Bell (2015) indicated that the principal aim of cross-sectional survey research designs revolves around the collection of quantifiable data concerning the same research subject with a view of detecting any patterns of association within the secondary data and the triangulation of key findings from multiple sources.



The research study utilizes content analysis and desk based research strategy with an inductive approach by first distilling the appropriate work in the literature and then constructing thresholds to assess the efficacy of the a IoT and Cloud Based HealthCare system.

The dependence on literature as major form of data extraction has its pros and cons, when using the content analysis technique; such challenges are identified by Kalenga (2015), which has been considered throughout the study as benchmark to objectify the information.

The aim of the method used is to precipitate qualitative information under a theme followed for the analysis. The quality of the published reports is also assessed based on the standards followed in the industry.













Chapter 4: End-to-End Architecture for enabling Artificial Intelligence through stream analytics

The end to end solution of the system comprises of the five architectural layers as shown and explained in table below. However, our main focus is on the “Real-Time Analytics” Layer, where the machine learning algorithm for the acquired streams of data are implemented on the data coming from the IoT hubs from heart beat monitors and other devices attached to the patients. For the purposes of predictive analytics the acquired data is first cleansed and then compared with historical logs of events recorded from multiple data sources. The algorithm is defined in a way such that the system is self-learning and self-optimizing.

Table 1: Architecture layers used for the system model.


The real-time event processing reference architecture is dependent on a number of different Azure PaaS capabilities to support end-to-end analytics scenarios. The following diagram and subsequent sections describe the various stages of the data flow lifecycle within this architecture, as well as the components required to support each of these stages, and the interaction between each of these components. The roles that each of the services illustrated above can support tends to blur across the layer boundaries. For example, SQL Database provides both a storage capability as well as a query engine for returning records to a user. Also, Event Hubs provides storage for a pre-defined period of time. This diagram is intended to illustrate each of the services being mapped to their primary role within this architecture.



The reference architecture used for the study is selected for building scalable and secure solutions that are device-centric which encapsulates the processing required for conducting predictive analysis and integrating the cloud infrastructure as required by the system proposed. The solution can be implemented on public, hybrid or private cloud as per the requirements of the system. The architecture enables a smooth flow of information between the connected health monitoring devices and line-of-business assets to the cloud infrastructure for the processing, analysis, control and business process integration. Figure 1 provides a high-level conceptual architecture which can be used as a reference model for the study and can be modified to serve specific identified purposes. The architecture is also aimed at catering the distributed nature of sensory instruments through a centralized cloud based processing system. The gateway is used to connect the device which has dedicated processing capabilities; it facilitates bidirectional communication with the back-end system and provides endpoints for the connected devices. The basic backend comprises of components for data collection, business logic, visualization, analytics, device discovery and registration.



Our main concern for the health monitoring system is to define the analytics layer for the system which shall be capable of streaming the live data and applying machine learning algorithms using third party data sources for enabling the system for artificial intelligence in disease discovery and prevention diagnosis. For this system the most appropriate module for streaming and analysing the data is the Azure Stream Analytics. After streaming the data the system utilizes the cortana studio for the implementation of the machine learning algorithm and the newly structured data is managed and stored using a combination of Azure Blob storage and SQL database.

There are two technologies that support the real-time analytics layer of this event processing reference architecture. Azure Stream Analytics acts as the linchpin of this reference architecture by providing the engine capable of temporal analytics over moving data streams. The second technology in the real-time analytics layer is Azure Machine Learning, which is a predictive analytics service capable of consuming either a single record made up of multiple columns (sometimes referred to in the context of machine learning as “features”) via request / response API, or consuming a file for asynchronous batch scoring. Given the real-time scope of this document, the role of Azure Machine Learning will be based on the request / response method only.

Additionally, Apache Storm is positioned in this diagram as a reference as to where it aligns to Azure Stream Analytics. Each of these platforms support similar event processing scenarios, however there are a number of differences in their supported languages and interoperation capabilities.

The figure below provides the reference architecture on which the system shall be based. The main module of concern for this study is the “Analytics and Machine Learning” component of the data processing, analysis and management section.

Figure 3:  Cloud based IoT solution Reference Architecture.

The technical details of the data acquisition devices will be provided and the embedded systems used as well as the sensing devices will be elaborated.



Stream Analytics:

Azure Stream Analytics ships with a pre-built adapter for Event Hubs, which significantly reduces the time to market for delivering solutions since developers do not need to develop their own custom adapters for the streaming engine. Additionally, and perhaps more significant for enabling accelerated solution deployment, Azure Stream Analytics provides a SQL-like language that allows database developers with existing SQL skills to transition to this platform very quickly. This language is very similar to T-SQL, which is the primary database language for SQL Server, however it contains a superset of functions that support temporal operations such as applying sliding, hopping or tumbling time windows to the event stream.

Azure Stream Analytics supports two different types of inputs, either stream data or reference data, and two different input data sources, either Event Hubs or files from Blob Storage. Only Blob Storage is supported for consuming reference data into the stream. Event streams can be consumed from either Blob Storage (in which case all events already exist in a file) or from Event Hubs whereby data is arriving in real-time. In the instance that data is being consumed from Blob Storage, the file being referenced must contain a timestamp record in a supported time

format for enabling Stream Analytics to perform temporal analysis. Whilst it is possible to stream

a file without a timestamp field that would not enable temporal queries, the value of the

Solution is somewhat diminished unless there is a specific scenario that requires this

Functionality. The following diagram represents the different data ingestion options for Stream


Machine Learning:

Azure Machine Learning provides a cloud based platform for mining data to identify trends and patterns across large scale information sets. By “learning” from historical events and trends, machine learning can publish a model that can be leveraged for determining predictions in real-time based on incoming event data. Combining the real-time rules based processing capabilities of Azure Stream Analytics with the real-time predictive analytics capabilities of Azure Machine Learning can help businesses rapidly deploy highly scalable data solutions to support complex information challenges.

Machine learning models that have been developed, trained and tested, are deployed as web services that can be called from applications. A request to a model is made asynchronously, either by submitting a single record containing a number of columns, or by submitting a file that contains a number of records to be scored. For the purposes of real-time analytics, the single record request / response method is used for submitting an incoming event to the machine learning model.

One of the considerations for constructing a single record feature set to submit to machine learning models is that the incoming event stream may not contain all of the data required for the model schema. For example, telemetry that is transmitted from equipment may contain data about the performance characteristics, but it will likely not contain all of the attributes of that equipment as it is repetitive and will bloat the data transmission. Equipment attributes, such as age may be highly influential features in a real-time machine learning model and therefore need to be appended to the record prior to calling the model.

The following diagram illustrates how Azure Machine Learning can make use of streaming data to enable real-time prediction.

The diagram above does not make an assumption on which action the application will perform

once the prediction has been received back from the machine learning service. There are a

number of possibilities here, one of which may be that the application calls a notification service

(Azure notifications are beyond the scope of this document) to send an alert to a device. Other

alternatives may include writing the prediction result to a SQL database for consumption into

reporting analytics, or initiating a broader automated workflow.

A single solution may contain multiple models that need to be trained for predicting different

outcomes based on the same or similar input data. Whilst this diagram depicts a single machine

learning model, an actual implementation may comprise multiple models being called from a

single orchestrating application, including scenarios where the response from one model may

result in the triggering of another model request in a sequential manner.


The Microsoft Azure Machine Learning Studio:


The sample data to test the algorithm is used for the cancer detection system. For the coding and scripting R script is used in the MS Azure Machine Learning Studio.


The Appendix 3 shows the complete R script implementation from the start till the end of the process for cancer prediction using cleansed data set for training and testing. The implementation of the system is conducted using the jupyter server. The AzureML package is installed by default on the jupyter server.

First of all the data set is loaded through the data storage layer. In order to make the system intuitive and intelligent the algorithm first trains the system. The clusters acquired with results are splited into the positives and negatives; which are further splited into halves for training and testing purposes. The further identified features of the cancer patients are also added individually into the dataset. The figure below defines the process of splitting and cleaning the data to be used for training and testing.


The data samples are then joined with the additional features column to make four independent data sets: Positive training examples, positive test examples, negative training examples, negative test examples. After that the test samples and training samples for negative and positive sets of data are joined.

The final cleansed training data set is then acquired for further analysis. The data is further grouped into bins and columns are selected to execute the following R script for training the model.

# replicate positive examples 93 times


dataset <- maml.mapInputPort(1)


data.set <- dataset[dataset[,1]==-1,]

pos <- dataset[dataset[,1]==1,]

for (i in 1:93) data.set <- rbind(data.set,pos)

row.names(data.set) <- NULL






Two individual data sets are then trained: firstly the dataset with edited metadata and secondly the data set without the editing of meta data.

After that the four data sets as shown in figure below are trained and scored.

Training and scoring models:


The models are then trained in four different sections using the “Train Model” module of the Microsoft Azure Machine learning studio. Then the patients IDs are used to test the trained model. The result output is executed using the following R script: Given in the appendix 1.


Training and evaluation of regression models is then conducted to check which regression model is most appropriate for the acquired data set. The following four regression models are used to train the data: Decision Forest regression, Boosted Decision tree regression, Poisson Regression, and Neural network regression. The rows for the scores of the models are then added and the following Rscript is used for the execution of the results for declaration of best model to be utilized: Given in Appendix 2.





Heart Disease Prediction:

A similar heart disease prediction training model is implemented in the MS Azure Machine Learning studio and is published. The diagram below gives the demonstration of algorithm.

Figure 2: Heart Disease Prediction Alorithm.(Published at:

The regulatory and ethical requirements revolve around the data handling procedures deployed in the proposed technological solution. Major stakeholder is the Data Protection Agency, responsible for ensuring the validity of the system in the light of Data Protection Act 1998 (Sophie et al. 2012). The proposed technological solution for the fall detection system deploys sensor nodes for data collection to monitor activity of the patient continuously. The data acquired shall be transmitted to the base station for further analysis and to be used by the stakeholders for assessment and retrieval. The stored data for retrieval include the patient’s complete medical information along with all the confidential details required for analysis by health professional. The system shall utilize the localization details of the patient at all times; which might be a source of unethical intervention in the personal lives of the patients if the data is not handled professionally by the stakeholders. According to the proposed technological solution for the system, second party (base station correspondents) shall be given complete access to the patient’s medical data  and third party (relatives) shall be given limited amount of access to patient’s data (Yi et al. 2014). For ensuring the alignment of the system with the rules and regulations applicable, following procedures and regulations are considered (Sophie et al. 2012):

  • APMS contract: arrangements made under section 83(2) of the 2006 Act for the provision of primary medical services.
  • The oversight mechanism of the National Health Service Commissioning Board.
  • Confidentiality and disclosure of information: General Medical Services (GMS) Code of Practice.
  • General Medical Council (2009) Confidentiality Code of Practice.

Given the vulnerability of the data acquired at the base station servers, it is very important that complete network security architecture shall be implemented; with cyber security layers deployed at multiple levels (Boric-Lubecke et al. 2014). A complete network security infrastructure shall be designed to ensure the security and integrity of the data, and make the system resistant to third party intrusions.





Chapter 5: System Requirements

  • The Client, the Customer and Other Stakeholders

The Client: Hospitals

The Customer: Elderly persons above age of 65 years and their families.

Other Stakeholders: Doctors, Nurses, Carers, Investors, Emergency Response correspondents, Data Protection Agency, and Medical Association.


  • Users of the Product

Hospital Personnel: Doctors, Nurses, Carers, Correspondents and Management.

Elderly Persons

Relatives of the Elderly



Project Constraints

  • Requirements Constraints

The project poses many technological and ergonomic constraints; due to the extent of requirements and limitations, posed by the required flexibility of application.

  • The system shall be reliable at multiple levels to increase the level of trust for the stakeholders: consistency in fall detection, accuracy of unusual behaviour detection, accuracy of coordinates acquisition using GPS, appropriate battery life for 18 hours of non-stop operation, availability of assistance at all times, and consistency in extent of access of information for every stakeholder (Salem et al. 2013).
  • The system shall use complete cloud computing solution for databases and real time networking which shall be secure and free of latency in VOIP.
  • The system must adapt and evolve to minimize errors and enhance accuracy by integrating machine learning algorithms.
  • User interface must be simple and easy to use; satisfying usability demands for the elderly.
  • The wearable device must be easy to handle and lightweight to ensure that the elderly persons can comfortably operate it, avoiding non-usage and refused-usage behaviours. The design must be non-intrusive and maximally discrete (Patel et al. 2012).
  • The collected data must be organized and available in a way to make sense for professional and non-professional users (Marschollek et al. 2012).
  • The number of nodes for end to end data acquisition and communication must be minimized (Lim et al. 2015).

Technological constraints:

  • The device must continuously monitor movement and localization of the user to perform real-time and transparent evaluations which shall be used to make alarm triggering decision using threshold algorithm.
  • The system shall take into account constraints and limitations posed by wearable device and smartphone integration, due to limited battery life (Lim et al. 2015).
  • The size of the sensor node must be minimized to ensure non-intrusive usage (Patel et al. 2012).
  • Data must be available for updates and real-time evaluation continuously; so, the system shall make use of internet of things for connecting servers to the user nodes (Salem et al. 2013).
  • Naming conventions and definitions

GPS: Global Positioning System

GPRS: General Packet Radio Service

WSN: Wireless Sensor Network

WBAN: Wireless Body Area Network

Sensor Node: The wearable device equipped with accelerometer and gyroscope.

Data Acquisition: Acquiring data and tagging it with appropriate relevance for understanding and retrieval.

Data Fusion: Using data from different sources to predict the outcome.

Medical Database: The medical history of the patient maintained at the base centre.

Internet-of-things: A web of sensors and computer devices connected to each other through internet.

Machine Learning: Intelligent agent algorithm which performs changes in the thresholds of the system to enhance the effectiveness of the detection system automatically.

Wearable device: The necklace designed to be worn by the elderly patient, which transmits sensor data to the wireless transmitter and mobile phone.


  • Relevant facts and assumptions
  1. The most important assumption is about the usability of wearable device: The assumption is ‘normal usage’ behaviour of the elderly towards wearable device, that is, the person is wearing an emergency wireless transmitter like wrist watch, a necklace or a pendant permanently during the whole day time.
  2. The correspondents at the base centre are available at all times for monitoring activity of the elderly patients and providing assistance in case of fall event detection or abnormal movement detection.


Functional Requirements:

  • The scope of the work

The scope of work for the system extends from customized hardware designs to the integration of hardware with customized software at multiple levels. Data-acquisition and Data-fusion shall be performed using real-time evaluations in the database servers with the aid of threshold and machine learning algorithms using cloud platform of Microsoft Azure. The system shall use GPS, GPRS, WSN and WBAN technology to transmit collected sensor data to the database servers and website in real-time (He et al. 2012). The updated information regarding activity and patient condition must be available at an online platform; which shall be accessed by all stakeholders for analysis and updates.

Azure’s predictive analytics services, including Machine Learning, Cortana Analytics and Stream Analytics, will be used for health intelligence. Doctors will be able to make smarter decisions, improve customer service and uncover new business possibilities from structured, unstructured and streaming Internet of Things data.

  • For IOT data handeling and control the system shall use: Azure IOT Hub, Stream Analytics
  • For networking requirements of the system: Azure DNS, CDN, ExpressRoute and Virtual Network.
  • For customer interaction: Business SAAS apps



Some major requirements of the system include:

  1. Fall Detection System:
  • Automatic emergency response infrastructure:

Alarm in case the elderly person comes out of the care home and garden area;

Alarm in case of no movement for a set period of time; and

Alarm after detection of sudden acceleration and fall.

  • Manual emergency alarm and communication:

The elderly person using the system shall be able to trigger an emergency alarm voluntarily, bypassing the automatic system. Furthermore, the elderly person shall have an option of communicating with the correspondent at base centre at any time.

  • False Alarm Designation:

He elderly person shall be able to respond immediately to a false alarm by signalling a false alarm designation immediately (Salem et al. 2013).

  1. User Localization:

The system must be capable of tracking the position of the elderly at all times. GPS technology along with GSM shall be used to localize the user.

  1. Real-time and transparent evaluation of the movement:

The movement pattern must be evaluated by the system in real-time using the kinematical activity data-with the use of accelerometers; the movement divided into three types: low, medium and high. Abnormal activity must be detected and an alarm shall be triggered in response (Yi et al. 2014).

  1. Bidirectional Voice communication:

Both the correspondent and the elderly person shall have the flexibility of communicating with each other at all times.

  1. Access to battery level information:

The battery level of the devices must be visible to all stakeholders through online platform (Razzaque and Dobson 2014).

  1. Machine Learning Algorithms for data mining and decision making using real time sensor data

The system shall be capable of learning new behaviours and adjusting itself in real time using data server storage and real time data acquisition. Situations such as false alarm trigger and non-detection of fall shall be utilized by smart agents to enhance the algorithm (Kangas et al. 2014).

  1. Indoor data transmission
  2. Outdoor Data transmission

Outdoor data transmission shall be conducted using android phone application connected to the wearable device via Bluetooth connection as mechanised by Casilari et al. (2015)

  • The scope of the product

The product must be able to use a small web of internet of things in the form of sensor nodes and connect it to the internet for real-time data acquisition and decision making. The product shall be based on multiple hardware devices to be used by the elderly, a data acquisition and decision making system to trigger emergency response, an online platform to be used by professionals and relatives of the elderly to analyse and update information, and a data-server to store and interpret information online. Different UCI Software shall be designed: for, the sensor node device; the smart phone; the online platform; MySQL databases; and emergency response centres (Yi et al. 2014).


Some of the major requirements for the product include:

  • The product must be cost effective.
  • The product must be easy to handle and discrete.
  • The product must be adaptable.
  • The product must have the flexibility of customization.
  • The product must maximally aid in management.
  • The product must have minimal power requirements.
  • The product must have long battery life.


  • Functional and data requirements

The data should be acquired and transmitted in real time with adequate amount of transparency to the online database and the system shall use Microsoft Azure (open, flexible and enterprise grade platform). The online database then can be accessed by the stakeholders for analysis and updates. The updates must be available at online and call centre platforms, which shall use cloud services and data storage facilities.


The collected data shall go through a process of data-fusion to integrate sensory information from different nodes. The movement information along with the sensory data from the wearable device shall be available in real-time for the analysis of professionals (Yi et al. 2014). The data collected shall be used by intelligent software agent to predict if an unusual activity has occurred. The unusual activity shall be categorized into three different ways by the algorithm automatically at the base station (call centre) to predict which emergency response alarm must be triggered (Akbar et al. 2015).


The patient information available at the online platform must have different level of access for different stakeholders, depending on their privileges defined by the management.


Firstly, the data shall only be available online to the registered users and each registered user shall be assigned a different privilege by the management.


Secondly, the relatives of the patient should be able to access the real time sensory data being collected by the sensors but shall not be allowed to access all the comments delineated by the professionals.

Thirdly, an activity report shall be generated once a day and its comparison shall be made by normal activity to sense any possible dangers. This shall be done using data-mining algorithms implemented on the acquired data.


Fourthly, the medical professional shall be given complete access to the history and reports generated by patient’s activity. They shall be allowed to amend the data and add their comments and diagnosis online.


Non-functional Requirements:

  • Look and Feel Requirements

The product shall utilize a minimalist design and interface standard. The wearable sensor node shall be made as attractive as possible with minimum number of buttons so as to make it less noticeable.


  • Usability and Humanity Requirements

The wearable device must be easy to handle and lightweight to ensure that the elderly can comfortably operate it-avoiding non-usage and refused-usage behaviours. The design must be non-intrusive and maximally discrete.


The system shall be dependable enough to induce a sense of security and confidence in the patients which can possibly add to its worth and users will deviate less from normal usage.


  • Performance Requirements

The system shall be reliable at multiple levels, to increase the level of trust for the stakeholders: consistency in fall detection, accuracy of unusual behaviour detection, accuracy of coordinates acquisition using GPS, appropriate battery life for 18 hours of non-stop operation, availability of assistance at all times, and consistency in extent of access of information for every stakeholder (Akbar et al.2015).




  • Operational and environmental Requirements

The system shall be capable of operating in all sorts of environment. To ensure the operations in variable environments the wearable device must be water proof, shock resistant, shall be capable of operating in extreme temperatures.

  • Maintainability and Support Requirements

A base station shall be mechanized with a purpose of collecting data through wireless transmission, storing data in the data base, using data-mining and data fusion algorithms to trigger emergency response, and a call centre to respond to the emergency needs and communicate with the patient (Kapadia et al. 2015).


The medical professional along with other authorized users of the online platform must have access to the battery level of the fall detection sensor node. The system must trigger an alarm in case the battery needs recharging.


  • Security Requirements

The data acquired through sensor nodes must only be available to users with authorized access and strict measures must be taken to ensure security of the data at the base station. Complete security requirements protocol is established in Part 2.


  • Cultural Requirements

Given that the sensor node is a wearable device; it must be designed such that it is not detectable while the user interacts with other people, so that peers do not pass judgmental remarks.


  • Legal Requirements

UK Data Protection Rules and Regulations are applicable for the security of personal information of the patients.

The medical professional using the data and providing services shall be fully qualified for the jobs and should have a professional education.

For detailed legal and ethical requirements kindly see Part 2.

Project Issues:

  • Open Issues

The issues are all covered in the requirements constraints section of the template.

  • Off-the-Shelf Solutions


The indoor and outdoor operation:

To enable the system to collect data and transmit it in real time at any place the patient might be, a triple node bidirectional transmission mechanism shall be designed.


For indoor application a wireless transmission device is connected to the wearable sensor.


For outdoor fall detection a mobile phone shall be used, connected to the wearable sensor device through Bluetooth technology (Casilari et al. 2015).


The data collected shall be retained and stored at the call centre to be available for data mining in order to detect unusual activity.


Proposed Technological solution:

The proposed technological solution of the system is mechanised taking into close consideration the requirements mentioned in the Volere Template.

The system shall utilize the same infrastructure for fall detection. However, for indoor operations system shall utilize the PERS (personal emergency response system) infrastructure which consists of a wearable device and a base station used for data transmission to the base centre and mobile communication (Casilari et al. 2015).

Personal Emergency Response System shall comprise of three nodes:

  • The wearable device
  • The wireless transmitter remotely connected to the database centre using PERS systems.
  • A mobile device used for communication with the base centre in case of emergency outdoors- the use of built-in tri-accelerometer of smart phone along with GPS data transmission system shall be utilized for fall detection (He et al. 2012).


The data collected shall be analysed in real time using threshold algorithms to initiate emergency response infrastructure as proposed by Phu et al. (2015). The data fusion shall be done over the internet through utilizing MySQL database integrating nodes of the system in a web of IoT. The issues as delineated by Kapadia et al. (2015) shall be considered to implement end to end medical informatics.

For the implementation of the hardware infrastructure an INGA wireless sensor node shall be used (Kapadia et al. 2015). The node comprises of an accelerometer, a gyroscope and a barometric pressure sensor. Accelerometer shall transmit the primary data acquired to detect abnormal behaviour. The wireless sensor node is equipped with an IEEE 802.15.4 compatible radio transceiver, which is widely used in the area of Wireless Sensor Networks (WSNs) and Wireless Body Area Networks (WBANs) (Lim et al. 2015).


















Diagram 1: Operating Environment: Indoor and Outdoor

Figure 1: SysML requirements diagram: Operational environment.



Diagram 2: Type of alarm depending on the movement:

Figure 2: SysML requirements diagram: Alarm.








  • Use Case 2:


Health professional analyses and updates patient information from the online platform.

Name: Update Medical reports through online platform.

Identifier: BC2



Primary Actor: Medical Professionals

Supporting Actor: Online Reception Software Application



  1. The medical professional has authorization of access and can amend the reports of patients.
  2. The system is running online with database maintained at the call centre.


Flow of events:

  1. The health professional access the online reception software application using web browser.
  2. The system prompts with the user-interface asking for log-in information. The health professional enters his user name and password.
  3. The system directs the professional to the available database according to his/her privilege of amendment and provides access to reports of the patients generated in the past. The system saves the activity log, tagged with recent activity of the user which can be retrieved at the call centre by Age Concerns management.
  4. The health professional uses the options on the interface to access the reports of the patient under observation. The system prompts with search for the patient number to access relevant data. The system interface also guides the professional to analyse real-time data received from the movement and fall detection devices.
  5. The health professional enters the unique patient ID in the search box.
  6. The system compares the privileges allowed to the user for the specific patient number.
  7. The system directs the user into the profile of the patient, from where the user can access previously generated reports and amend them accordingly.
  8. The user selects the report to be studied, using the date of the report generated.
  9. The user analyses the report and adds his suggestions and diagnosis along with messages for different stakeholders.
  10. The system tags the generated data with the name and ID of the health professional. It updates the database immediately and sends a notification to the stakeholders in their corresponding accounts. The carers and call centre professionals get the instructions generated by the system to act according following the advice of health professional.



The database is updated for every user in real-time.


Alternative Course:

  1. The unique patient ID entered by the health professional does not allows privileged access to the user, to study and amend the details.

5.1. System shows a warning message indicating that the user was denied access to the desired information and the user must contact the management team if he/she has detected an unusual behaviour.

5.2. User contacts the support team at the call centre, using Voice-Over-IP communication option at the website and delivers a request for the change in privilege for the patient.

Figure 4: Business Use Case Diagram 2.


Product Use Cases:

  • Use Case 1: Alarm activation system by the elderly without fall detection


The elderly patient feels a need to talk to the correspondent at the call centre bypassing the emergency alarm response.

Name: Update Medical reports through online platform.

Identifier: PC1



Primary Actor: Elderly Patient

Primary Actor: Correspondent

Supporting Actor: Call centre software



  1. The elderly person is wearing the necklace (sensor node): Normal Usage.
  2. The correspondent is available at the call centre.
  3. The wireless indoor link to the call centre is working and transmitting real time data to the call centre database.


Flow of events:

  1. The elderly patient detects an unusual health condition.
  2. The patient uses his/her pendant to press the button available for direct communication.
  3. The correspondent responds by contacting the patient through wireless GSM transmitter if he is indoors.
  4. The patient refuses to answer but presses the button on the pendant indicating emergency medical assistance required.
  5. The correspondent sends the message to the emergency response team which in turn show up at the door of elderly person using his localization information.
  6. The patient responds to the queries and receives the required emergency medical treatment from nurses.
  7. The patient answers the question if he/she wants doctor at his home or if he is willing to go to the care centre.



The emergency team reaches the location on time.


Alternative Course:

  1. The patient is outdoor on the road.

3.1. The data transferred to the base-centre is through the mobile phone.

  1. The patient answers the mobile phone and communicates his situation.

4.1. The correspondent analyses the situation and predicts the intensity of emergency.


Figure 3: Business Use Case Diagram 1.

Figure 3: Business Use Case Diagram 1.

Figure 5: Product Use Case Diagram 1.


  • Use Case 2: Information access by the relatives



Relatives of the elderly use the online platform to check the condition and activity of the elderly person.

Name: Relatives accessing online system platform.

Identifier: PC2



Primary Actor: Relatives of the elderly

Supporting Actor: online system platform



  1. The relative of elderly person has authorization of access to the information of the patient.
  2. The online system is updated from the base centre.

Flow of events:

  1. The relative of the elderly person uses the website URL to access his/her authorized account
  2. The website opens the login page and asks for the login information of the person.
  3. The user enters the username and password assigned by the management.
  4. The user selects the option for checking the patient’s medical condition.
  5. The user analyses the information and prints the data to keep the record to him.
  6. The user chooses the option to check comments by the medical professional.
  7. The user adds queries to the message board which will be answered by the health professional.
  8. The user logs out of the account.



The relative uses the information to serve the elderly better.


Alternative Course:

  1. The relative does not have authorized access to the patient’s information anymore.
  • The log-in page requests the relative to talk to call centre for help.


Figure 6: Product Use Case Diagram 2.


Adapt Case: Enhancing data analysis and decision making for fall detection using machine learning algorithms


Adaptation of better algorithm for unusual behaviour detection using false alarm data and the data acquired in case a fall occurred but the system failed to trigger the alarm.


Name: Use machine learning intelligent agent to enhance the reliability of the system.

Identifier: AC1



Primary Actor: Intelligent software agent

Primary Actor: Elderly person

Supporting Actor: Call centre software



  1. The elderly person experienced a fall but the alarm was not triggered at the base centre.
  2. The elderly person did not experience a fall but a false alarm was triggered at the base centre.


Flow of events:

  1. The elderly patient experiences an unusual health condition and experiences a fall.
  2. The system uses the data but does not comprehend the behaviour as a fall event.
  3. The elderly person uses the bypassing mechanism, to trigger a manual alarm.
  4. The intelligent agent stores the data acquired before the manual alarm to add to the data base as an example of a fall event.
  5. The intelligent agent adjusts the algorithm to adapt to the latest example data stored for fall event.


Alternative Course:

  1. The elderly patient does not experience an unusual health condition or a fall and a false alarm is triggered at the base centre.
  2. The system uses the data and comprehends the behaviour as a fall event.
  3. The elderly person uses false alarm button to indicate that an actual fall did not occurred.
  4. The intelligent agent stores the data acquired before the false alarm to add to the data base as an example of false fall event.
  5. The intelligent agent adjusts the algorithm to adapt to the latest example data stored for fall event.




Figure 7: Adapt Use Case Diagram.







Chapter 6: Conclusion and Recommendation

Ethical issues, Data Encryption and Privacy

Encryption shall be done over all communication channels from sensor node transmission to the PERS transmitter as well as Internet, ISP-based connections etc. Confidential data shall be kept encrypted on the workstations and data base station (Al Ameen et al. 2012). Strict change controls shall be used. Field level file change history shall be maintained. Digital signatures of creator and checker shall be acquired in real time.

Some important considerations for network and data security infrastructure include (Kapadia et al. 2015):

  • Firewalls-internet connection
  • Integrated smart card access control at the data base station
  • Encryption-Application specific
  • Database Security-Proprietary, DBMS-specific, RBAC (role based access control).
  • Authentication- User ID and password based with limited smart card pilots


To ensure ethical disclosure of patient’s medical data the following issues shall be taken into consideration:

  • The patient shall be informed of the data usage purposes: for direct clinical care (assessment by health professionals); and for secondary uses (access by the relatives). As according to the requirements online database shall be maintained for frequent access by the stakeholder to patient’s medical data. The patient shall be given access to online database website to change the privilege of access for any second party under the UK National Health Service (NHS), ‘The confidentiality and Disclosure of Information Directive 2013’ (2013).
  • The health professionals shall take into consideration the Data Protection Act 1998 and consider each patient as a unique case study. The online base station website shall delineate the implied rules from the Act to ensure that the doctors consider them before upholding any sort of data usage. Data Protection Act 1998 legislates the following:
  1. Medical data shall be discarded after its necessary usage is completed.
  2. The data required for different purpose shall be disclosed differently, to minimize disclosure amount.
  3. The stakeholders must ensure safety and security in data handling and storage.
  4. Information shall be retained about the place of data storage; the oversight mechanism shall assure the implementation of security and contractual agreements.
  5. A written assurance shall be provided to the stakeholder to outline the type and intensity of data disclosure, with a surety that the data will not be disclosed to a third party.
  • The secondary usage shall be maximally restricted to disclosure of effectively anonymised, pseudonymised or aggregated data. Such as the correspondent at the base station shall not be given access to patient’s personal information such as name, ID or picture. The emergency infrastructure shall use a unique code name provided to each patient. Section 251 of the NHS Act 2006 provides statutory basis for patients to restrict the release of medical data to a second or third party disclosure.
  • The stakeholders shall retain a key to code data and convert it from anonymised to specified form as advised by General Medical Council (2009) Confidentiality, GMC, London p.30.
  • The express consent of the patient is not required under the ‘Confidentiality and disclosure of information Directions 2013’, which delineates statutory basis for bypassing consent where it is not practical to hide data tagging; such as, disclosures relating to financial and management arrangements of NHS (regulations 2004).
  • A formal data sharing protocol shall be published on the base centre website.
  • The health professional shall have an approval of a written protocol delineating the structure of their trials for the patient as instructed by the report ‘Staff care: how to engage staff in the NHS and why it matters’ published by The Point of Care Foundation (2014). The correspondent shall stick to asking about specified and approved set of questions when contacting the elderly in case of emergency. All clinical trials shall be authorised by ethics committees working under NHS as part of the Health Research Authority’s National Research Ethics Service (NRES).






“Understanding Workflow in Telehealth Video Visits: Observations from the IDEATel Project.” Journal of Biomedical Informatics 42 (4): 581–592.

Adibi, S. 2012. “Link Technologies and BlackBerry Mobile Health (mHealth) Solutions: A Review.” IEEE Transactions on Information Technology in Biomedicine 16 (4): 586–597.

Alemdar, H., and C. Ersoy. 2010. “Wireless Sensor Networks for Healthcare: A Survey.” Computer Networks 54 (15): 2688–2710.

Atzori, L., A. Iera, and G. Morabito. 2010. “The Internet of Things: A survey.” Computer Networks 54 (15): 2787–2805.

B.-H. Yang, S. Rhee, and H. H. Asada, “A twenty-four hour tele-nursing system using a ring sensor,” in Proc. IEEE Int. Conf. Robot. Autom., vol. 1. May 1998, pp. 387392.

Baek, H. J., G. S. Chung, K. K. Kim, and K. S. Park. 2012. “A Smart Health Monitoring Chair for Nonintrusive Measurement of Biological Signals.” IEEE Transactions on Information Technology in Biomedicine 16 (1): 150–158.

Bryman, A. and Bell, E., 2015. Business research methods. Oxford University Press, USA

Bsoul, M., H. Minn, and L. Tamil. 2011. “Apnea MedAssist: Real-time Sleep Apnea Monitor Using Single-Lead ECG.” IEEE Transactions on Information Technology in Biomedicine 15 (3): 416–427.

  1. C. Chan, C. W. Chen, W. C. Chou, Y. L. Ho, Y. H. Lin, and H. P. Ma, “Live demonstration: A mobile ECG healthcare platform,” in Proc. IEEE Biomed. Circuits Syst. Conf. (BioCAS), Nov. 2012, pp. 8787.

Carroll, R., R. Cnossen, M. Schnell, and D. Simons. 2007. “Continua: An Interoperable Personal Healthcare Ecosystem.” IEEE Pervasive Computing 6 (4): 90–94.

Cegarra-Navarro, J.-G., and M. T.Sánchez-Polo. 2010. “Implementing Telemedicine Through e-listening in Hospital-in-the-Home Units.” International Journal of Information Management 30 (6): 552–558.

Constantinescu, L., J. Kim, and D. Feng. 2012. “SparkMed: A Framework for Dynamic Integration of Multimedia Medical Data into Distributed m-Health Systems.” IEEE Transactions on Information Technology in Biomedicine 16 (1): 40–52.

Corchado, J. M., J. Bajo, D. I. Tapia, and A. Abraham. 2010. “Using Heterogeneous Wireless Sensor Networks in a Telemonitoring System for Healthcare.” IEEE Transactions on Information Technology in Biomedicine 14 (2): 234–240.

Costa, Â., J. C. Castillo, P. Novais, A. Fernández-Caballero, and R. Simoes. 2012. “Sensor-Driven Agenda for Intelligent Home Care of the Elderly.” Expert Systems with Applications 39 (15): 12192–12204.

  1. J. Lee, M. K. Park, and J.-H. Lee, “Height adjustable multi-legged Giant Yardwalker for variable presence,” in Proc. IEEE Int. Conf. Adv. Intell. Mechatron. (AIM) Jul. 2015, pp. 104109.

De Capua, C., A. Meduri, and R. Morello. 2010. “A Smart ECG Measurement System Based on Web-Service-Oriented Architecture for Telemedicine Applications.” IEEE Transactions on Instrumentation and Measurement 59 (10): 2530–2538.

Delmastro, F. 2012. “Pervasive Communications in Healthcare.” Computer Communications 35 (11): 1284–1295.

Department of Economic and Social Affairs of the United Nations. 2009. World Population Ageing 2009. New York: United Nations.

Dolan, B. 2011. “10 Reasons Why Google Health Failed.” MobiHealthnews by Chester Street Publishing Inc., 27 Jun.

Domingo, M. C. 2012. “An Overview of the Internet of Things for People with Disabilities.” Journal of Network and Computer Applications 35 (2): 584–596.

Duan, L., and L. Xu. 2012. “Business Intelligence for Enterprise Systems: A Survey.” IEEE Transactions on Industrial Informatics 8 (3): 679–687.

Duan, L., W. N. Street, and E. Xu. 2011. “Healthcare Information Systems: Data Mining Methods in the Creation of a Clinical Recommender System.” Enterprise Information Systems 5 (2): 169–181.

Escarrabill, J., T. Martí, and E. Torrente. 2011. “Good morning, Doctor Google.” Revista Portuguesa de Pneumologia (English Edition) 17 (4): 177–181.

European Commission Information Society. 2008. Internet of Things in 2020: A Roadmap for the Future.

European Commission Information Society. 2009. Internet of Things Strategic Research Roadmap 2009.

Eurostat. 2008. “The Life of Women and Men in Europe: A Statistical Portrait.” In Statistical Book, edited by Luis del Barrio, 15–16. Luxembourg: Eurostat.

  1. Axisa, A. Dittmar, and G. Delhomme, “Smart clothes for the monitoring in real-time and conditions of physiological, emotional and sensory reactions of human,” in Proc. 25th Annu. Int. Conf. IEEE-Eng.-Med.-Biol.- Soc., Cancún, Mexico, Sep. 2003.

Gao, J., Z. Pang, Q. Chen, and L. Zheng. 2010. “Interactive Packaging Solutions Based on RFID Technology and Controlled Delamination Material.” In Proceedings of IEEE International Conference on RFID, Orlando, FL, April 14–16, 158–165. New York: IEEE. doi:10.1109/ RFID.2010.5467270.

  1. C. Koh and G. Tan, “Data mining applications in healthcare,” J. Healthcare Inf. Manage., vol. 19, no. 2, p. 65, 2011.
  2. Rifaï, S. Mohammed, and A. Mellouk, “A brief synthesis of QoS-QoE methodologies,” in Proc. 10th Int. Symp. Program. Syst. (ISPS), Apr. 2011, pp. 3238.
  3. S. Alqahtani and P. Sant, “A multi-cloud approach for secure data storage on smart device,” in Proc. IEEE 6th Int. Conf. Digital Inf. Commun. Technol. Appl. (DICTAP), Jul. 2016, pp. 6369.

He, W., and L. D. Xu. 2012. “Integration of Distributed Enterprise Applications: A Survey.” IEEE Transactions on Industrial Informatics, doi:10.1109/TII.2012.2189221.

Hossain, M., and D. Ahmed. 2012. “Virtual Caregiver: An Ambient-aware Elderly Monitoring System.” IEEE Transactions on Information Technology in Biomedicine. doi:10.1109/ TITB.2012.2203313.

  1. Drago, M. Mellia, M. M. Munafo, A. Sperotto, R. Sadre, and A. Pras, “Inside dropbox: Understanding personal cloud storage services,” in Proc. ACM Conf. Internet Meas. Conf., Nov. 2012, pp. 481494.

Istepanaian, R. S. H., and Y.-T. Zhang. 2012. “Guest Editorial Introduction to the Special Section: 4G Health—The Long-Term Evolution of m-Health.” IEEE Transactions on Information Technology in Biomedicine 16 (1): 1–5.

ITU. 2005. The Internet of Things-Executive Summary. internetofthings.

Jeong, S., C.-H. Youn, E. B. Shim, M. Kim, Y. M. Cho, and L. Peng. 2012. “An Integrated Healthcare System for Personalized Chronic Disease Care in Home–Hospital Environments.” IEEE Transactions on Information Technology in Biomedicine 16 (4): 572–585.

Junnila, S., H. Kailanto, J. Merilahti, A.-M. Vainio, A. Vehkaoja, M. Zakrzewski, and J. Hyttinen. 2010. “Wireless, Multipurpose In-Home Health Monitoring Platform: Two Case Trials.” IEEE

  1. Dolui, S. Mukherjee, S. K. Datta, andV. Rajamani, “ReTiHA: Real time health advice and action using smart devices,” in Proc. Int. Conf. Control Instrum. Commun. Comput. Technol. (ICCICCT), Jul. 2014, pp. 979984.
  2. Lin, W. Wang, X. Wang, W. Ji, and J. Wan, “QoE-driven spectrum assignment for 5G wireless networks using SDR,” IEEE Wireless Commun., vol. 22, no. 6, pp. 4855, Dec. 2015.

KALENGA, R. C. (2015). Adjusting western research techniques to accommodate research in the indigenous realm. Indilinga African Journal of Indigenous Knowledge Systems.14, 1-            11.

Kim, T.-W., and H.-C. Kim. 2012. “A Healthcare System as a Service in the Context of Vital Signs: Proposing a Framework for Realizing a Model.” Computers & Mathematics with Applications 64 (5): 1324–1332.

Klasnja, P., and W. Pratt. 2012. “Healthcare in the Pocket: Mapping the Space of Mobile-Phone Health Interventions.” Journal of Biomedical Informatics 45 (1): 184–198.

Koop, C. E., R. Mosher, L. Kun, J. Geiling, E. Grigg, S. Long, C. Macedonia, R. Merrell, R. Satava, and J. Rosen. 2008. “Future Delivery of Health Care: Cybercare.” IEEE Engineering in Medicine and Biology Magazine 27 (6): 29–38.

Kratina I.S., Zamfirescu C.M., Zamfirescu C.M., Trepanier K., & Marques L. (2015).Survey questionnaires and graphs. Electronic Journal of Statistics.9, 2202-2254.

Li, C. Wu, L. Guo, C. H. Lee, Y. Guo, “Wiki-Health: A Big Data Platform for Health Sensor Data Management”, IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014.

Li, L., Li, X., Hueiwang, A. J., Dayanand, N., Thomas, A., & Maria, F. (2008). Creation of environmental health information system for public health service: A pilot study. Information Systems Frontiers.

  1. Chen, “NDNC-BAN: Supporting rich media healthcare services via named data networking in cloud-assisted wireless body area networks,” Inf. Sci., vol. 284, pp. 142156, Nov. 2014.
  2. Chen, J. Wan, S. Gonzalez, X. Liao, and V. C. M. Leung, “A survey of recent developments in home M2M networks,” IEEE Commun. Surveys Tuts., vol. 16, no. 1, pp. 98114, Mar. 2014.
  3. Chen, S. Mao, and Y. Liu, “Big data: A survey,” Mobile Netw. Appl., vol. 19, no. 2, pp. 171209, Apr. 2014.
  4. Chen, Y. Hao, C. F. Lai, D. Wu, Y. Li, and K. Hwang, “Opportunistic task scheduling over co-located clouds in mobile environment,” IEEE Trans. Serv. Comput., to be published, doi: 10.1109/TSC.2016.2589247.
  5. Chen, Y. Hao, M. Qiu, J. Song, D.Wu, and I. Humar, “Mobility-aware caching and computation ofoading in 5G ultra-dense cellular networks,” Sensors, vol. 16, no. 7, p. 974, 2016.
  6. Chen, Y. Hao, Y. Li, C.-F. Lai, and D. Wu, “On the computation ofoading at ad hoc cloudlet: Architecture and service modes,” IEEE Commun. Mag., vol. 53, no. 6, pp. 1824, Jun. 2015.
  7. Chen, Y. Ma, J. Song, C. F. Lai, and B. Hu, “Smart clothing: Connecting human with clouds and big data for sustainable health monitoring,” Mobile Netw. Appl., vol. 21, no. 5, pp. 825845, 2016, doi: 10.1007/s11036-016-0745-1.
  8. Chen, Y. Qian, S. Mao, W. Tang, and X. Yang, “Softwarede ned mobile networks security,” Mobile Netw. Appl., vol. 21, no. 5, pp. 729743, 2016, doi: 10.1007/s11036-015-0665-5.
  9. Chen, Y. Zhang, L. Hu, T. Taleb, and Z. Sheng, “Cloud-based wireless network: Virtualized, recongurable, smart wireless network to enable 5G technologies,” Mobile Netw. Appl., vol. 20, no. 6, pp. 704712, Dec. 2015.
  10. Chen, Y. Zhang, Y. Li, M. M. Hassan, and A. Alamri, “AIWAC: Affective interaction through wearable computing and cloud technology,” IEEE Wireless Commun., vol. 22, no. 1, pp. 2027, Feb. 2015.
  11. Chen, Y. Zhang, Y. Li, S. Mao, and V. C. M. Leung, “EMC: Emotionaware mobile cloud computing in 5G,” IEEE Netw., vol. 29, no. 2, pp. 3238, Mar./Apr. 2015.

Microsoft Corporation. (2016). Azure IoT reference architecture. Retrieved January 12, 2017, from

  1. S. Pandian et al., “Smart vest: Wearable multi-parameter remote physiological monitoring system,” Med. Eng. Phys., vol. 30, no. 4, pp. 466477, 2008.

Pang, Z., Q. Chen, and L. Zheng. 2009. “A Pervasive and Preventive Healthcare Solution for Medication Noncompliance and Daily Monitoring.” In Proceedings of 2nd International

Pang, Z., Q. Chen, and L. Zheng. 2012. “Value creation, Sensor Portfolio and Information Fusion of Internet-of-Things Solutions for Food Supply Chains.” Information Systems Frontiers. doi:10.1007/s10796-012-9374–9.

Pang, Z., Q. Chen, L. Zheng, and E. Dubrova. 2013b. “An In-home Medication Management Solution Based on Intelligent Package and Ubiquitous Sensing.” In Proceedings of International Conference on Advanced Communications Technology, Global IT Research Institute (GiRI), Republic of Korea, January 27–30.

Paulraj, D., S. Swamynathan, and M. Madhaiyan. 2012. “Process Model-Based Atomic Service Discovery and Composition of Composite Semantic Web Services Using Web Ontology Language for Services (OWL-S).” Enterprise Information Systems. doi:10.1080/ 17517575.2011.654265.

Pawar, P., V. Jones, B. F. V. Beijnum, and H. Hermens. 2012. “A Framework for the Comparison of Mobile Patient Monitoring Systems.” Journal of Biomedical Informatics 45 (3): 544–556.

Plaza, I., L. Martína, S. Martinb, and C. Medranoa. 2011. “Mobile Applications in an Aging Society: Status and Trends.” Journal of Systems and Software 84 (11): 1977–1988.

Poon, C. C. Y., and Y.-T. Zhang. 2008. “Perspectives on High Technologies for Low-Cost Healthcare.” IEEE Engineering in Medicine and Biology Magazine 27 (5): 42–47.

  1. Liu, Y. Ma, M. Alhussein, Y. Zhang, and L. Peng, “Green data center with IoT sensing and cloud-assisted smart temperature control system,” Comput. Netw., vol. 101, pp. 104112, Jun. 2016.

R W. Picard, Affective Computing. London, U.K.: MIT Press, 1997.

Rodrigues, J. J. P. C., O. R. E. Pereira, and P. A. C. S. Neves. 2011. “Biofeedback Data Visualization for Body Sensor Networks.” Journal of Network and Computer Applications 34 (1): 151–158.

  1. Hu and J. Tan, “BioLogger: A wireless physiological sensing and logging system with applications in poultry science,” in Proc. IEEE Annu. Int. Conf. Eng. Med. Biol. Soc., Sep. 2009, pp. 48284831.
  2. Tivatansakul, M. Ohkura, S. Puangpontip, and T. Achalakul, “Emotional healthcare system: Emotion detection by facial expressions using Japanese database,” in Proc. 6th Comput. Sci. Electron. Eng. Conf. (CEEC), Sep. 2014, pp. 4146.

Serbanati, L. D., F. L. Ricci, G. Mercurio, and A. Vasilateanu. 2011. “Steps towards a Digital Health Ecosystem.” Journal of Biomedical Informatics 44 (4): 621–636.

Sinderen, M., and J. P. A. Almeida. 2011. “Empowering Enterprises through Next-Generation Enterprise Computing.” Enterprise Information Systems 5 (1): 1–8.

Symposium on Applied Sciences in Biomedical and Communication Technologies (ISABEL2009), 1–6. doi:10.1109/ISABEL.2009.5373681.

Tartarisco, G., G. Baldus, D. Corda, R. Raso, A. Arnao, M. Ferro, A. Gaggioli, and G. Pioggia. 2011. “Personal Health System Architecture for Stress Monitoring and Support to Clinical Decisions.” Computer Communications 35 (11): 1296–1305.

Teng, X., Y.-T. Zhang, C. C. Y. Poon, and P. Bonato. 2008. “Wearable Medical Systems for p-Health.” IEEE Reviews in Biomedical Engineering 1: 62–74.

Tollmar, K., F. Bentley, and C. Viedma. 2012. “Mobile Health Mashups: Making Sense of Multiple Streams of Wellbeing and Contextual Data for Presentation on a Mobile Device.” In Proceedings of the 6th International Conference on Pervasive Computing Technologies for Healthcare (PervasiveHealth), San Diego, CA, May 21–24, 65–72. New York: IEEE.

Transactions on Information Technology in Biomedicine 14 (2): 447–455. Kaufman, D. R., J. Pevzner, M. Rodriguez, J. J. Cimino, S. Ebner, L. Fields, V. Moreno, et al. 2009.

  1. Anliker et al., “AMON: A wearable multiparameter medical monitoring and alert system,” IEEE Trans. Inf. Technol. Biomed., vol. 8, no. 4, pp. 415427, Dec. 2004.

Vincent, G. K., and V. A. Velkof. 2010. The Older Population in the United States: 2010 to 2050: Population Estimates and Projections. New York: United States Census Bureau.

Wang, H., Y. Shi, and W. Xi. 2012. “Development of Medicine Management System for RFIDBased Intelligent Medicine Cabinet.” WAP Conference Series: Education and Social Science Review 2012 (1): 1–4.

World Health Organization. 2011. mhealth: New Horizon for Health Through Mobile Technologies. Global Observatory for e-Health Services, vol. 3. Geneva: WHO.

  1. Fan, C. He, Y. Cai, and Y. Li, “HCloud: A novel application-oriented cloud platform for preventive healthcare,” in Proc. IEEE 4th Int. Conf. Cloud Comput. Technol. Sci. (CloudCom), Dec. 2012, pp. 705710

Xu, E., M. Wermus, and D. B. Bauman. 2011. “Development of an Integrated Medical Supply Information System.” Enterprise Information Systems 5 (3): 385–399.

Xu, L. 2011. “Enterprise Systems: State-of-the-Art and Future Trends.” IEEE Transactions on Industrial Informatics 7 (4): 630–640.

  1. Zhang, “GroRec: A group-centric intelligent recommender system integrating social, mobile and big data technologies,” IEEE Trans. Serv. Comput., vol. 9, no. 5, pp. 786795, Sep./Oct. 2016, doi: 10.1109/TSC. 2016.2592520.
  2. Zhang, D. Zhang, M. M. Hassan, A. Alamri, and L. M. Peng, “CADRE: Cloud-assisted drug recommendation service for online pharmacies,” Mobile Netw. Appl., vol. 20, no. 3, pp. 348355, 2015.
  3. Zhang, M. Chen, D. Huang, D. Wu, and Y. Li, “iDoctor: Personalized and professionalized medical recommendations based on hybrid matrix factorization,” Future Generat. Comput. Syst., vol. 66, pp. 3035, Jan. 2017.
  4. Zhang, M. Chen, S. Mao, L. Hu, and V. Leung, “CAP: Community activity prediction based on big data analysis,” IEEE Netw., vol. 28, no. 4, pp. 5257, Jul./Aug. 2014.
  5. Zhang, M. Qiu, C. W. Tsai, M. M. Hassan, and A. Alamri, “Health-CPS: Healthcare cyber-physical system assisted by cloud and big data,” IEEE Syst. J., 2015, doi: 10.1109/JSYST.2015.2460747.

Yang, G., L. Xie, M. Mantysalo, J. Chen, H. Tenhunen, and L. R. Zheng. 2012. “Bio-Patch Design and Implementation Based on a Low-Power System-on-Chip and Paper-based Inkjet Printing Technology.” IEEE Transactions on Information Technology in Biomedicine. doi:10.1109/ TITB.2012.2204437.

Yin, J., H. Chen, S. Deng, Z. Wu, and C. Pu. 2009. “A Dependable ESB Framework for Service Integration.” IEEE Internet Computing 13 (2): 26–34.

















Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s