Contextual activity based Healthcare Internet of Things, Services, and People (HIoTSP): An architectural framework for healthcare monitoring using wearable sensors

Abstract

Healthcare industry is gaining a lot of attention due to its technological advancement and the miniaturization in the form of wearable sensors. IoT-driven healthcare industry has mainly focused on the integration of sensors rather than the integration of services and people. Nonetheless, the framework for IoT-driven healthcare applications are significantly lacking. In addition, the use of semantics for ontological reasoning and the integration of mobile applications into a single framework have also been ignored in many existing studies. This work presents the implementation of Healthcare Internet of Things, Services, and People (HIoTSP) framework using wearable sensor technology. It is designed to achieve the low-cost (consumer devices), the easiness to use (interface), and the pervasiveness (wearable sensors) for healthcare monitoring along with the integration of services and agents like doctors or caregivers. The proposed framework provides the functionalities for data acquisition from wearable sensors, contextual activity recognition, automatic selection of services and applications, user interface, and value-added services such as alert generation, recommendations, and visualization. We used the publicly available dataset, PAMAP2 which is a physical activity monitoring dataset, for deriving the contextual activity. Fall and stress detection services are implemented as case studies for validating the realization of the proposed framework. Experimental analysis shows that we achieve, 87.16% accuracy for low-level contextual activities and 84.06%–86.36% for high-level contextual activities, respectively. We also achieved 91.68% and 82.93% accuracies for fall and stress detection services, respectively. The result is quite satisfactory, considering that all these services have been implemented using pervasive devices with the low-sampling rate. The real-time applicability of the proposed framework is validated by performing the response time analysis for both the services. We also provide suggestions to cope with the scalability and security issues using the HIoTSP framework and we intend to implement those suggestions in our future work.

Introduction

The integration of wireless communication, sensor, and information technology has opened new paradigms in the field of wellbeing and healthcare management. It enables the quality of life and extends the autonomous living of individuals at home. Today, embedded sensors can measure a wide range of physical parameters with the minimal complexity. This phenomenon becomes realistic due to their low power consumption and cheaper production cost. These embedded sensors can be integrated in small wearable devices like watches, pendants, and clothes. Moreover, they can also be applied to everyday objects or places, such as appliances, furniture, or the like, for automation purposes. Internet of Things (IoT) has radically changed the way the human interacts with things, i.e. sensors, objects, services, and applications. These entities provide continuous data streaming over a protracted period of time, based on its characteristics. The acquired data from these diversified set of sensors can be analyzed and used to enable a service or process an action based on the user's request [1], [2], [3]. Since IoT is not bounded to any particular field, it has been applied to a wide range of services [1], [3], [4], [5].

IoT has profoundly changed the way of healthcare service delivery, and as a result, physicians recommend the use of health-related applications to patients [6]. Many applications are available which take input from the users about their daily routines and dietary habits [7], [8] to report their health condition, but IoT driven healthcare can benefit from the automatic collection of information through internal and external sensors for detecting anomalous conditions. These sensors refer to the built-in sensors in smartphones [9] and wearable sensors like smartwatches [10]. IoT has been extensively studied for smart homes. The related services are based on the changes in sensor readings such as humidity, temperature, ambient light sensors, and so forth. The changes in the sensor measurements are used to control the home appliances and enhance the quality of living without user intervention. Researchers are still trying to assess the complete spectrum of IoT driven healthcare systems. There is a dire need for powerful tools that can acquire data from multiple sensor modalities such as psychological, physiological, and inertial data to validate the autonomous healthcare monitoring system.

Some studies have stressed on the acquisition of multimodal data [11] whereas others focus on analyzing the provenance of the acquired data [12]. Healthcare IoT based services are mainly implemented with inertial measurement units (IMUs) for activity recognition, event/process detection using object/location sensors, and physical state monitoring using physiological sensors [13], [14]. Most of the works have been performed using event/process detection with object sensors and activity recognition. These works are specifically proposed for either inhouse patients, elders or people with specific disease which requires a proper setup or infrastructure to be deployed, hence, trading off with mobility and easiness of such systems. Moreover, the use of physiological sensors has not been exploited to its full potential yet in the field of healthcare driven IoT services. It is apparent that more usage of sensors results in wider scope of the health-related services. In this regard, we propose a framework which can combine multiple sensor modalities to provide the healthcare services in an accurate and an efficient way for all users. The use of wearable sensors in this framework allows the system to be location-independent or infrastructure-less system.

In general, IoT is a four-layered architecture with sensing, network/access, service and interface layer [1], [2], [3], [4], [5], [15], [16]. The sensing layer is responsible for the integration of different sensors, their connection to the physical world, and collection of data. The network/access layer provides the networking support and is responsible for the transfer of the data in wireless or wired networks. The service layer is managing and creating services as per the user's requirements. The interface layer is commonly used for presenting the analyzed data or the output of the desired services. The users or associated entities are able to interact with the IoT system using the interface layer. In general, IoT framework with specified layers is limited to some specific services but including another layer such as context recognizer can extend the scope of healthcare driven IoT and can be used for the automation of services. The important requirements for designing IoT based healthcare monitoring framework are summed up as follows [1], [2], [3], [4], [5], [15], [16]:

• Holistic solution: healthcare driven IoT can be used for physical fitness, safety, and health, as this solution encompasses everyone needs.

• Fusion of technologies: healthcare driven IoT can be integrated with a wide variety of sensors and can support multi-modal sensor platforms.

• Analytics and BigData: with recent advancement in cloud computing technology, multi-modal, multi-scale, and heterogeneous data can be easily pre-processed and analyzed in a reasonable time. This allows the applicability of real-time systems for healthcare services.

• Personalized services: analytics and big data techniques such as machine learning and recommendation systems expand the possibility for personalized healthcare treatments and services. It can also help early detection of anomalous conditions.

• Data for many time-lines: users can receive or access their past, present, and future predicted data anytime and anywhere.

• Telehealth features: the data and output from desired services can be shared with doctors in real-time. Moreover, doctors can monitor larger number of patients with the use of healthcare driven IoT systems.

In accordance with the requirements above, in this paper, we introduce the healthcare driven Internet of Things, services, and people (HIoTSP) framework in the context of wellbeing and healthcare using smartphone and wearable devices as shown in Fig. 1. HIoTSP is the interconnection of things (sensors), services (application and service triggering based on analysis) and people (caregivers or doctors) via the Internet for healthcare services [17]. This framework integrates the use of inertial, physiological, and location sensors for our context recognizer for deriving contextual activity to automate the service selection process. The sensor layer in HIoTSP is responsible for acquiring measurements from various sensor modalities. These sensor measurements are then transferred to the middleware via the access and communication layer which provide the means for interfacing. The middleware in the proposed framework is the smartphone which acts as a gateway for sending and receiving the information to and from the server layer. It can also perform some lightweight operations with the data stored in the buffer. The data received from the middleware is then accumulated in the storage of server layer. The server layer uses the data to recognize contextual activity to select the desired service(s), respectively. The data in compliant to the selected service is then passed to the service selection block for performing further analysis and decision making. This data is also used for summarizing and recording the event logs. The decision from the service selection is periodically checked by the middleware. As the decision is available, it is fetched by the middleware and immediately relayed to the notification center for alert generation and recommendations (the details of each of the blocks are provided in Section 3).

The proposed framework supports mobility as it can be used indoors as well as outdoors, low cost as it uses only wearable sensors and common devices such as smartphones, and user-friendliness as devices are unobtrusive, less complex, and easy to operate. The framework uses ontological reasoning or inference rules to select an application or service for analyzing, summarizing, and visualizing the data. The interface layer (notification center) in this framework explicitly deals with generating declarations based on the attained results. We have also implemented two health-related services i.e. fall and stress detection which are considered to be major anomalous conditions in the literature. The current implementation of the framework is geared towards research and small-scale field tests rather than towards the usage in practice with patients and caregivers. The contributions of the proposed work are summarized as follows:

• Integrated framework for healthcare driven IoT services that includes semantics (context recognizer) layer

• Combining multiple sensor modalities to automate the service selection as well as to cover wide range of health monitoring services.

• In-depth analysis for contextual activity recognition and related services such as fall and stress detection.

• Real-time implementation and validation of the proposed HIoTSP framework.

The rest of the paper is structured as follows, Section 2 highlights the existing studies and frameworks; Section 3 presents the proposed architecture of our framework. Section 4 provides the details for the implementation details of the framework and services. Section 5 explains the results from our implementation of framework modules and services. Finally, the discussion and conclusion are presented in Section 6.