1
Introduction

John Davies1, and Carolina Fortuna2

1British Telecommunications plc, Ipswich, UK

2Jozef Stefan Institute, Ljubljana, Slovenia

The physical world is becoming ever more closely connected to information systems as sensors and actuators are incorporated into a wide variety of physical objects - from highways to pacemakers to cattle to running shoes to factories - and then connected to the Internet via a range of wired and wireless networks. This is the Internet of Things (IoT) and it is already generating massive volumes of data. The result is that much richer information can be collected (in real time) and used by automated systems to provide actionable insight and to respond to changing contexts with appropriate intelligent actions. IoT has rapidly moved from the conceptual phase to widespread use in real-world applications in recent years.

The IoT will deliver significant innovation in many different areas, including future cities, transport, health and social care, manufacturing, and agriculture. Sensors can now be deployed at low cost to instrument the world to a far greater extent than has been possible before. There is increasing recognition of the potential value in opening up data resources so that they can be exploited more fully.

At the highest level, many of the IoT applications being considered appear similar - involving the collection of information from a range of sensors and other sources, interpreting this in a specific context, and then making better decisions that improve a behaviour or a process. For instance, smart watches or other types of wearable sensing devices are able to drive improvements in our behaviour toward a healthier daily routine. Merchandise tracking sensors can lead to a better understanding of supply chains and deliver optimization of costs and minimization of carbon footprints. IoT has a unique potential for automating and improving man-made systems and behaviours by enabling unprecedented understanding and insight. For example, IoT data enabled a recent comprehensive global study across 111 countries on the impact of physical activity variation and the built environment on health [1].

The IoT has been a recurrent theme among commentators since the term was coined in the late 1990s. The concept has evolved from early work on Radio Frequency Identifier (RFID) technology which represented a hardware related break-through that aimed to connect everyday objects to a network. This perhaps constituted the first wave of the IoT, which then developed beyond the initial hardware world innovation, and focused increasingly on developing new types of sensors and sensing materials, as well as on developing new communication technologies and protocols. As a result, a wide variety of new communication technologies emerged in the early years of the twenty-first century which were able to support the ubiquitous deployment of a wide variety of sensors. We refer to this as the second wave of IoT. In the last decade, the focus of IoT has shifted to data collection, processing and security aspects and this period is termed the third wave of IoT. This book focuses primarily on this most recent wave and covers all key aspects including data management, processing, and analytics as well as security, privacy and trust as depicted in Figure 1.1. Real-world examples are given that show the application of IoT technologies in a number of different sectors.

Figure 1.1 The IoT ecosystem.

1.1 Stakeholders in IoT Ecosystems

A number of different actors typically participate in any deployment of IoT technology and we will refer to this set of stakeholders and the relationships between them as the IoT ecosystem. Such stakeholders may play one or more different roles. These include sensor providers, connectivity providers, information providers, application developers, analytics service providers, platform providers, and end users of information and applications.

Information providers in IoT ecosystems are often owners of sensor deployments. The primary purpose of their sensors may be for their own use but they may choose to make some of their data available to others, either on a commercial basis, to meet their obligations (particularly for public sector organizations), or for the general good. Various data processing platforms may also be information providers, even if they are not directly associated with "Things." We refer to these as derived information providers; while not being the primary source of any information, they create value by combining data from multiple sources, transforming or applying various analytical techniques. These additive data sources could include: contextual (e.g. geographical, administrative) information; notifications of events such as traffic incidents and sporting fixtures; or, perhaps, rare events such as anomalies in a production process.

In efficient IoT ecosystems, information providers should be able to easily publish their services or data resources and advertise their availability via an easily accessible catalogue so that potential users can independently discover and assess their utility. This scenario is perhaps similar to the app stores that are commonplace today to make applications easily available. It is important to note that making data available should not imply relinquishing ownership rights; consequently, information providers also need the ability to define access controls, together with terms and conditions for use of the data they publish.

Platform providers have a key enabling role in the IoT ecosystem. They do not directly provide information or build dedicated services or applications but support stakeholders in other roles by providing a set of functionalities that all can use. This allows other participants in the ecosystem to focus on their own core activities and helps to accelerate innovation in the ecosystem. Platform providers may provide computing and storage infrastructure, as well as analytics services, which could include artificial intelligence (AI) capabilities such as summarization, enrichment, and reasoning.

Each platform provider will use specialist hardware and software tools and offer general-purpose frameworks that an end user can exploit to define their own workflows. For instance, an edge or cloud provider offers on demand compute and storage resources that can be configured and modified on demand by users. Certain platform providers, typically application domain experts, offer a more complete service, including consultancy services, to support end users who may not have the necessary systems, data science, or analytics expertise.

Application developers produce applications that process the available data within a specific context to produce actionable insight for end users. Application developers should be able to discover what data and platform resources are available to them, what the key features and costs of each resource are, and assess which ones meet the needs of the applications that they want to build. This includes both the information content of the resources and practical considerations for the resources, such as dependability (accuracy, availability, etc.), conditions of use, or commercial considerations.

End users participate in the ecosystem by using the information and applications that are made available to them by other stakeholders. The end users can be private persons or institutional decision makers. As the ultimate beneficiaries of the functionality provided by the other stakeholders, it is important that their experience is positive and the ecosystem delivers real value for them. An IoT ecosystem will not be sustainable without the trust of its end users.

For individual end users, participation in the ecosystem is generally via an application. Often this application will make use of information that is generated through their use of the application, for example the user's location is often used as a data source by applications on mobile phones. The situation where the individual is an information provider needs to be addressed with care, particularly where personally identifiable or potentially sensitive information may be involved. Open engagement with end users that ensures they are properly informed and understand that they are included in the ecosystem is essential.

1.2 Human and IoT Sensing, Reasoning, and Actuation: An Analogy

Along with IoT, artificial intelligence (AI) comprises an increasingly pervasive and important set of technologies. Recent years have seen significant advances in AI in a number of areas [2]. IoT and AI are inevitably interconnected, given the vast volumes of rich data generated by IoT and the ever-increasing capability of AI systems to analyze, extract insight, and make decisions from that data. Thus, any discussion of the role and impact of AI would be incomplete without consideration of the link to IoT and in this volume a number of chapters are included that discuss the role of AI in IoT systems.

The vision of the IoT is that digital systems can be given the ability to sense, process, and extract useful information and actionable insight from the world and respond to the environment accordingly (typically via actuation). From an AI and robotics perspective, we can make an analogy with human sensing/actuating capabilities and the five human senses that receive inputs from the external environment. These...