Keywords

2.1 A Brief History of Pervasive Computing

The healthcare Emerging technological change emphasized Researchers, to rethink about the more advanced and low cast technoogical methods and innovative systems like improvement in electronics technology (hardwrae devices) which are responsible to auto sensing capabilities (sensors), to measure all physical entities of real world and manipulate the data for further research studies. Similarly, telecommunication technologies avail high speed of Internet connectivity for traditional desktops, laptops, smart phone as well as embedded devices to ensure scalability, data transferring rates and connectivity of devices while sending data over network. Such enhancements in all afore-mentioned fields have opened door for researchers that innovated new dimension in the field of computer science called pervasive/ubiquitous computing. The words “pervasive” and “ubiquitous” are used interchangeably to describe the same concept.

Pervasive computing is the cutting-edge technology for the contemporary era. The pioneer of pervasive computing/ubiquitous computing was Mark Weiser [1], in PARC (Palo Alto Research Center) in America. Mark Wiser describes the concept of pervasive computing as “Inspired by the social scientists, philosophers and anthropologists at PARC, we have been trying to take a radical look at what computing and networking ought to be like. We believe that people live through their practices and tacit knowledge, so that the most powerful things are those that are effectively invisible in use. This is a challenge that affects all of computer sciences. Our preliminary approach: Activate the world. Provide hundreds of wireless computing devices per person per office of all scales (from 1″ displays to wall-sized). This has required new work in operating systems, user interfaces, networks, wireless, displays and many other areas. We call our work ‘ubiquitous computing.’ This is different from PDAs [personal digital assistants], Dynabooks or information at your fingertips. It is invisible, everywhere computing that does not live on a personal device of any sort but is in the woodwork everywhere”. Pervasive computing has changed the entire world as we perceive. Pervasive computing-enabled Internet of Things (IoT) allow the device to perform action while understanding the context. One of key features of pervasive computing is that it eliminated interaction of the user with application and performed all task on behalf of the user which made it as an invisible servant. These are the main features of pervasive computing according to its nature of working, context awareness, adaptation, scalability, and heterogeneity.

2.2 Main Features of Pervasive Computing

2.2.1 Context Awareness

According to Henricksen et al. [2], an application will be accomplished by reducing input from users and replacing it with a knowledge of context. Context awareness is a software component which exploits information such as activities in which the user is engaged, proximity, to other devices and services, location, time of day and weather conditions. This all information will change application behaviour according to users’ condition.

2.2.2 Adaptation

Adaptation is another feature of pervasive computing which is directly related to the user interface for pervasive computing environment that must be highly adaptable in order to respond to changes in the available input and output devices.

2.2.3 Scalability

In the pervasive computing , the factor of scalbility play an important role for healthcare fascilities such factors like time complexity, cost effectiveness, resources managment, hardware and software systems’ synchroization, these are very effective parameters for further detailed studies in the context.

2.2.4 Heterogeneity

Heterogeneity is related to the combination of different computing devices.

Ubiquitous computing provides two unique parameters anytime and anywhere for anyone which means the user can perform his or her task from anytime and anywhere by using one of the backbone services of pervasive computing called networking, which enable the user to control and perform task according to their need anytime and anywhere. Pervasive/ubiquitous computing covers a wide range of technologies to accomplish its invisible silent servant needs such as distributed and mobile computing, sensor networks, wireless sensor networks, human-computer interaction, artificial intelligence knowledge, etc. [3]. This technology has produced a great impact and brought several changes in every field of domain, because pervasive/ubiquitous computing means everywhere and anytime. This invisible servant work is accomplished by a small, tiny hardware called sensor which senses the physical phenomena and converts that data into the electrical form for further action. These sensors always have limited energy capability to live. The challenges of pervasive computing devices has been literatured by the researchers, they had identified, some are very importants like energey consumption, dat security& trust, data transmission & management and across the board heterogeneous capability. Therefore, there are lots of applications in each domain of life such as environmental monitoring (e.g. weather forecast, forest fire detection), natural disasters (e.g. flood detection and monitoring, earthquake), military (e.g. military equipment monitoring/surveillance, battle field management), business (e.g. product monitoring and tracking), and smart homes (e.g. monitoring and controlling home appliances) [4], and healthcare [5] (measure temperature, heart rate, blood pressure, blood and urine chemical level, breathing rate and volume, sugar level, and patient activities, etc.). Ubiquitous computing has brought great revolution especially in the medical field to change the way of traditional treatment from self-treatment. To automated treatment by using different sensors and actuators, monitoring devices are designed regarding the disease, which enable doctors to monitor each activity of patient such as patient name and location (room number, time, date) as well as to assist the patient from anywhere at anytime if emergency will occur and suggest specific medication for the patient to prevent the impact of disease as well as recover patient health. Many applications are designed in pervasive computing in the healthcare field to make patient treatment easier than traditional treatment. According to literature in traditional treatment, the writing error of doctors would be the cause of patient death [6], and medical error will be imposing substantial cost in national economy.

2.3 Pervasive in Healthcare

The pervasive computing has produced a great impact on every domain of life and has improved the working style of their traditional way atmost level by increasing their performance and as well as cost effcetiveness in the respective field. It is because it will save time and cost of the nation as well as individual person, and in emergency medical situation, medical support will be sent quickly and reduce death ratio of humans because pervasive computing automatically monitors the patient every activity and update doctors if any emergency occurred. Also, it provides real-time data in each second which is used to evaluate the health condition of patients who are suffering from critical health issue specially heart attack, diabetes, and cancer. Therefore, doctors and hospitals needed records of such patient(s) frequently to monitor their condition and also change the treatment if necessary. Accordingly, all fields of life have started integration of pervasive computing functionality.

2.4 Pervasive Applications in Healthcare

The pervasive computing in the context of application, there are many functionality attributes, taking more focussed on patients care, such as information security, automation, web services, monitoring etc. The following are several applications of pervasive computing in healthcare domain from different corporations:

  1. 1.

    Measure temperature: Use sensors to continuously measure the patient’s temperature to prevent the highest level of temperature that would be the cause of the patient’s death.

  2. 2.

    Heart rate: Likewise use pervasive computing to measure heart beat fluctuation in the form of an uninterrupted manner and patient heart rate to avoid serious and uncontrolled health condition, because the patient faces sudden and extreme critical health condition.

  3. 3.

    Blood pressure: Measuring blood pressure of each patient is a normal route task, but the major concern of this parameter is regarding the more critical condition of the patient with diabetes or those patients who are suffering from blood pressure. It is necessary to continuously measure this parameter of the patient to avoid the flow of blood in high-frequency rate that will burst the veins of the patient.

  4. 4.

    Blood and urine chemical level: Similarly, measure the blood as well as urine level of those patients who are suffering from the chemical level in blood as well as urine, if the urine chemical level crossed its highest level that will cause kidney stone. Automatically measure and control the urine and blood level to prevent critical condition of the patient.

  5. 5.

    Breathing rate and volume and sugar level: Pervasive computing is also used to monitor breathing rate and volume as well as the sugar level of the patient to recover from a health condition as soon as possible, but these all monitoring and measuring activities will be achieved without human interaction. Also ensure to ring alarm for doctors and nurses if any critical condition occurred.

  6. 6.

    Lung cancer: According to WHO reports, and its alarming that mostly cancer deaths observed due to smoking in the world. Smoking fog slowly damage the working capability of lungs and patient therefore, the key attributes shall be considered such as to monitor the lung cancer, recovery stage data and these data shall be analyzed in real time mode through sensors.

  7. 7.

    Patient activities: Patient activities can be classified into two types: one is auditor based and second is gesture based. Auditor activities are used to monitor the movement of patients whether he/she do the same activities which is prevented by the doctor or not indicating the attitude of patient towards instructions given by doctor, that must be followed by patient, other case all activities such as unconscious or partially hit patients.

Measuring heart rates of the patient in critical condition is important for doctors to analyse real-time improvements in patient’s health. Similarly nowadays people have acquired more stress which could be the cause of blood pressure. The record of heart patients, such as blood pressure, heart beat rate, temperature, weight, medicine and other related factors monitored for getting the analysis report of the patient recovery. Impure utilization of water would be the cause of kidney disease and block the urine veins. Likewise all these disease treatment will be done when doctor(s) attain real-time data of patient(s) and activities according to the disease that needs proper treatment. These applications of healthcare indicate the importance of pervasive computing in healthcare to provide a quick response to patient to save his/her life. Using pervasive computing to implement recognition of activity of daily life (ADL) using different classifiers and decision-making algorithms provide better healthcare and lifestyle management as well as exploit the machine learning algorithms to predict any disease and provide preliminary stage medication to the patient.

2.4.1 Security Features for Patients’ Data

In medical perspective the security feature is one of the key role factors about patient data, because the entire medicine will be prescribed based on the gathered data. Therefore, security of patient data is more critical because the entire decision is based on the gathered data. If the data is accurate and secure, then decision will be correct, and the doctor will prescribe the appropriate medicine for the patient. Otherwise doctors will make such erroneous decision about the patient’s treatment due to the gathered forgery-based data, which might be the cause of critical health condition of the patient or death. Therefore, secure confidential-based gathered data will make the doctors and hospitals gain confidence in taking decisions about patient treatment to recover from his/her health condition as soon as possible. Several security approaches have been designed and implemented for ensuring data integrity and confidentiality of patient data [7,8,9,10]. Because patient data are always important which are gathered by using body wearable network devices and using wireless sensor networks as an intermediary layer to transmit data over the network, WSNs are an open environment; therefore, it could be easy for attackers to apply any kind of attack (attacker may apply either passive attack which can see and analyse sensitive data but not able to modify it, and it discloses all confidential data, but active attacker can analyse as well as modify sensitive data and resend to the receiver) to modify the data [9]. Therefore, authentication and authorization of data along with integrity and confidential are more reliable and trustworthy data [4], which helps doctor(s) write proper and accurate medicine for the patient to save his/her life.

According to the working nature of pervasive computing technology, avaibility of resources and capability of monitoring through networks over internet invites many threats and attacksers. There are several types of attack, but a few of them are important. The common attacks are as follows:

  1. 1.

    DOS (Denial-of-Service) Attack

    The DoS attack was developed first time for testing of the network bandwidth capacity, but later this service attack is used to exploit the entire bandwidth capacity and prevent the access of authenticated user for their desired resources from the network for temporary or infinite time. Due to sending many requests over the network to keep the server busy as well as to consume complete bandwidth, DoS attack cannot be thwart through technical means alone [11]. In September 1996 the first-time attack was applied on Panix, the third-oldest ISP service provider in the world. This kind of attack is difficult to prevent through technical means alone because of how to identify which packets would be the cause of DoS attack.

  2. 2.

    DDOS (Distributed Denial-of-Service) Attack

    Distributed denial service attack is another type of DoS in this multiple systems that initiate attack to flood bandwidth or resource of target machine and ensure unavailability of resource from an authenticated user for temporary or intended time.

  3. 3.

    Man-in-the-Middle Attack

    Man-in-the-middle attack is the type of active attacker which intercepts between two communication parties either observing or modifying sensitive data.

  4. 4.

    Node Cloning Attack

    This attack is very difficult to find because all nodes which are cloned mean that they have compromised on security issue. An adversary physically captures a sensor node to extract all the secrete cryptographic from and reprogram it and again deploy that cloned node on the network to apply a variety of inside attacks on the network. This attack is applied on mobile or wireless sensor networks [12, 13].

  5. 5.

    Sniffing Attack

    This is another type of network attack which is common. In this attack adversary will capture the network traffic and steal important information such as password. A file also compromises the confidentiality of the security aspect [14].

  6. 6.

    Packet Drop Attack

    This is another type of attack which is applied by adversary on the wireless sensor network which is deployed in hostel and unattended environment which easily becomes the victim of an adversary. The malicious node will drop packet to prevent its further propagation. Therefore, an adversary does not drop each packet, but it drops selective packets [15,16,17].

There are many pervasive computing systems that have been designed as well as implemented in healthcare system. A few of the systems have been described in this section, such as Intelligent Heart Disease Prediction System (IHDPS) for detecting heart disease proposed by different classifiers like Decision Tree Naïve Bayes and Neural Network to predict and identity the significant impact of attributes of heart disease, and many other propose approaches have been used to achieve the healthcare aim and also many applications regarding to healthcare [18]. Another one is Ubiquitous Healthcare Information System (UHIS) which enables the patient to access healthcare services anywhere and anytime for further reading to see [19].

2.4.2 Lifecycle of Pervasive Healthcare Data

The concept of this lifecycle with respect to pervasive healthcare data may be described in the following six major domains. This is a proposed lifecycle model for pervasive healthcare data:

  1. 1.

    Control systems (sensors, databases, management systems)

  2. 2.

    Tool (apps, data mining, monitoring, communication systems, reporting)

  3. 3.

    Extract transform load (ETL) tool (integration, filtrations of data)

  4. 4.

    Dynamic selections (data schemas, labelling of data)

  5. 5.

    Patient’s activities-monitoring

  6. 6.

    Patients’ behaviour assessment/mood readings

As per the need of healthcare system, we have enlisted the above stages for our proposed model for healthcare which is known as lifecycle of pervasive healthcare system. Because without these stages, any proposed system would not be implemented as per the nature of pervasive computing. Therefore, all the above list of stages fall in the healthcare lifecycle of pervasive computing which is represented in pictorial form in Fig. 2.1. Each stage consists of certain components and tasks to perform several actions and send that data as input for the next stages which also perform a necessary task and then send that output as input to another stage, etc.

  1. 1.

    Control System (Sensors, Databases, Management System)

    The control system depends on four core objects as given above. The sensor is a tiny device which is responsible for sensing physical phenomenon and sending that gathered data either for processing purpose or storage purpose which is entirely based on purposed system and how it will exploit gathered data. Similarly database is a tool which is capable of storing data for future requirement according to the need. Because database data is also important when the same patient will come back again in the hospital, his/her demography will help the doctors take quick steps to save his/her life. The whole control is being managed by the administrators by helping in supporting the staff further which is classified into:

    1. (a)

      Devices (pervasive/sensor-based) manager

    2. (b)

      Database manager

    3. (c)

      Host network manager

    4. (d)

      Communication systems (satellite) manager

    5. (e)

      Application software manager

  2. 2.

    Tool (Apps, Data Mining, Monitoring, Communication Systems, Reporting)

    Tool is the second stage of healthcare lifecycle which consists of six major components that enable the healthcare lifecycle to perform meaningful task. App is indicating any application that is responsible for receiving data from any sensor, and it is front-end for the doctors to analyse fluctuation of receiving the data which gives an entire improvement of clue of the patient’s health. This fluctuation data exploits data mining functionality because data mining extracts data on pattern base and produces knowledge discovery report (which is called meaningful data extracted from the gathered data) [20,21,22]. Monitoring is also front-end for doctors, nurses, and hospital paramedical staff while they continuously monitor the patient’s health. The communication system is the soul of the gathering data because it works as the intermediary layer to increase scalability of network and to ensure data transmission over the network. The bandwidth also plays a key role in communication system which affect the data transmission rates if data transmission rates are slow which directly emit an impact on system performance due to high latency. Reporting is the last component of tool it generates the final report that will be used for making decision. It is necessary for reporting that data must be extracted by data mining techniques.

    The concept of tools is generic, but here we have limitation to the following:

    1. (a)

      Application of front-end tools

    2. (b)

      Database system and visualization tools

    3. (c)

      Network tools

    4. (d)

      Communication tools

    5. (e)

      Operating system tools

    6. (f)

      Sensor device controlling and data extraction tools

    7. (g)

      Data warehousing tools

    8. (h)

      ETL tools

  3. 3.

    Extract Transform Load (ETL) Tool (Integration, Filtrations of Data)

    This is used to integrate the ETL tools with application. As ETL will be integrated with the existing application tool/systems to keep the connection with source data for getting the information or fields/labels, this works like a supporting module with middleware application for extracting the required data, transforming it according to the design of data warehouse, and loading it in the suggested mart if required.

    The ETL tool is available as open source with SQL Server 2012 and the latest versions with special wizard: SQL Server Integration Services (SSIS). Before getting in the process, data sources may be classified in these MySQL, TEXT, XML, Excel, CSV, and SQ-DB types. These sources may be live or offline data sources, according to the nature of the data provider.

  4. 4.

    Dynamic Selections (Data Schemas, Labelling of Data)

    In this stage, the system deals with dynamics of all data, including the design of data whether star schemas, snowflake, or galaxy, OLAP framework in synchronous data, data warehousing, and further data mart design. This shall need labelling of data to extract and filter the data attributes of all dynamics like personal profiles, medical staff, and management systems either retaining the same labelling or changing with the suggested labels to better understand the dynamics of data.

  5. 5.

    Patient Activities (Monitoring (Auditory and Gestured))

    The fifth stage of healthcare lifecycle deals with patient activities to avoid the prohibited things as well as activities. For example, if the doctor restricted the patient not to smoke, then it is necessary for him to monitor the patient’s activities as he/she might be smoking. Auditory data deals with voice communication of patient, whereas gestured data belongs to the patient’s movement regarding eye flipping, hand raising, finger movement, etc. All these data are also important in the health perspective.

  6. 6.

    Patient Behaviour Assessment/Mood Reading

    Patient assessment is an important factor in the healthcare domain. On the bases of assessment, doctors and nurses predict improvement in the patient’s health. This health improvement is entirely based on the patient moods because if the patient gets well, then he/she feels happy and feverish. Mood or behaviour may be classified into two types: one may be positive mood and second may be negative mood. The measurement of these moods is divided into at least two parameters according to the situation and timestamp. Using these parameters, we can calculate whether the patient is in which mood, either in positive or negative mood.

Fig. 2.1
figure 1

Lifecycle of pervasive healthcare data

Full size image

2.4.3 Classification of Pervasive Healthcare Data

The classification of pervasive healthcare data is a very complex subject; however, this can be classified into different types of data: text, images, audios, and videos oriented in online transactional processing (OLTP). These types further can be classified in some domain areas where it has correlation with each type of pervasive healthcare data. The following are the important pervasive healthcare data specifications:

  1. 1.

    Personal profile (patients, doctors, nursing staff, account/management staff, pharma staff, IT staff)

  2. 2.

    Hospital management systems (monitoring, management, control, evaluation systems, updating)

  3. 3.

    Nursing facilities (availability of required nursing facilities, staff, resources, etc.)

  4. 4.

    Medicine (online complete data of stock and valid information)

  5. 5.

    OPD data (recording OPD data and mining the same)

  6. 6.

    Hospitalized data (patient’s hospital history data storage)

  7. 7.

    Home arrangement data (home service monitoring and analysis data history)

  8. 8.

    Testing data (keep all records of patients’ test data for ready reference)

2.4.4 Dynamics of Pervasive Healthcare Data

The following key components may be involved in the pervasive healthcare data dynamics:

  1. 1.

    Patients

  2. 2.

    Doctors

  3. 3.

    Nursing staff

  4. 4.

    Hospital management

  5. 5.

    Applications/sensor tools

  6. 6.

    Laboratories/testing tools

  7. 7.

    Medicine

  8. 8.

    Sensor systems/machines

  9. 9.

    Diseases classifications

  10. 10.

    Networks/communication technologies

  11. 11.

    Cloud technologies/data mining/data warehousing

2.4.5 Dynamics Labelling Process for Healthcare Data Designing

The data is a real asset for any kind of knowledge, in the pervasive healthcare data, specially depending on how accurate and complete the data is. This data shall be integrated through valid labels for analysing and monitoring the record of all dynamics, as patients, doctors, etc. This large data set is to be generated round the clock, and this activity is monitored by sensor machines automatically. The valid data collection is a very complex job, as well as very expensive, too. Authors Cruz-Sandoval et al. [23] presented two approaches in monitoring the patient’s activity through online systems for data extraction:

  1. 1.

    Measuring the burden on the user (gesture-oriented activity), who is performing and annotating the activity

  2. 2.

    Counting the lack of accuracy due to the user (auditory data-oriented) labelling the data (minutes/hours every activity)

The labelling may vary into labels like behaviour, timestamp, etc. It also seems like a segmentation and labelling [24]. The main reason for the labelling of data is to conclude and summarize the facts for knowledge extraction. There are many approaches to get labelled data by using algorithms and techniques. This labelled data can be processed and utilized in many ways for live running sessions as well as future assessment and analysis. The pervasive data by nature is a kind of live data such as online transactional processing (OLTP) data to save the precious lives of the patients. It may either in a synchronous or asynchronous manner depending on the situations. This data may further be utilized and warehoused for future use, and it also can be mined for research and artificial intelligence purposes. In the context of pervasive data, new emerging techniques are used to label the data. Some are mentioned here to understand the labelling concept and its importance in the field of pervasive healthcare data.

2.4.6 Pervasive Healthcare Data Labelling Techniques

The following approaches/techniques can be referred to data labelling with respect to one of the dynamics of pervasive healthcare data [25,26,27,28,29,30,31,32,33]:

  1. 1.

    Temporal labelling (synchronous/asynchronous)

  2. 2.

    Annotator (observer)

  3. 3.

    Scenario

  4. 4.

    Tool annotation mechanism

  5. 5.

    Survey-based tool annotation mechanism

The data is the backbone of all pervasive systems, so this is a complex dynamic and shall be addressed and mainly cared for to handle the devices as well as persons engaged with the systems. The pervasive healthcare data is a very sensitive and big asset that needs a lot of attention regarding security, integration, consistency, transparency, and being available for all stakeholders. We tried to focus our attention on healthcare data and its behaviours relating to the systems integrated with data and communication systems. The model we suggested in this study shall be effective to compose the theory about pervasive computing in detail. In this model, six components are involved to carry on the complete processing of pervasive healthcare data. Missing of at least one would damage the structure of pervasive computing theory. The patient and sensor-based device technologies are key factors and play an important role in the pervasive technologies. Patient care is the major activity, with supporting factors like medical staff, IT support team, and sensor devices as well as the main entities in the pervasive health data. The control systems become the starting layer in the process, being physical and logical architecture. It depends on this process and after this next shall be the tool process to handle the situations and make taking care of the patients possible. This suggested model loop will carry on the whole pervasive system smoothly and effectively till it is achieved.