Search

The cost of delivering social programmes (health, education, national security etc.) in all OECD countries is rising to an unsustainable level. The Internet of Things and increased connectivity are increasingly being seen by public and private sector organisations as a way of reducing costs and improving service provision. As ICT progresses from existing in standalone devices to all connected devices (IoT) it will provide new opportunities to improve and integrate service provision. However, increasingly the physical existence of humans and their physical environments will become a key part of smart systems and the services they provide. Before this can happen, there is first a need to better understand how physical things can be codified thereby allowing them to be included into a system. Secondly, there is a need to integrate and harmonise all things (appliances, devices, fixtures/fittings, building fabric/components, community services etc.). Thirdly, the social outcomes of that integration need to be understood, and this calls for a new nomenclature, a ‘social language’. This research seeks to address these three challenges by discussing the prioritisation of the needs of an individual and characterising all things in terms of the service they provide. It then proposes how these things might be ‘connected’ to form a mesh network that encompasses both current physical, and future virtual, service-delivery mechanisms. The findings indicate that there is potential to codify things by their services in a meaningful way. It is also clear that this approach would offer much greater analysis of service delivery and be able to optimise that delivery in a more integrated and connected world. To develop the work further requires considerable programming capability and capacity. However, while this research has involved largely theoretical scenarios, it has also developed several online ‘tools’ driven by the need to provide proof-of-concept models to support research as well as potentially generate income. The objective to develop a ‘social language’ was acknowledged from the start as being extremely complex, in essence, trying to codify and systemise individual and community needs—even culture itself. However, it is a starting point for others to fulfil its objective in a practical way. The prize is to truly understand the social value of buildings and things in a smarter future and thereby improve wellbeing and social equity for all.

A set of systems are considered where there are multiple agents simultaneously active within real-time a environment. These systems are termed SMART systems and are found in domains as diverse as office administration, process control and aviation. Such systems place particular demands on agents that are not present in non-SMART systems. Actions may be time constrained in two ways: i) an action appropriate at one time may not be appropriate at another; ii) the time available for a required action may approximate to the time in which an agent is able to perform that action. In order to be able to function in such environments agents, both human and computer, must be aware of time constraints and the actions necessary to ensure that they do not compromise system goals.

Verification and validation are very important in the design of any system. For safety-critical systems, especially, the validation level is crucial. In this work, developing a robust procedure for designing smart, safe, secure systems is our eventual goal. There are different methods for verification and validation, but formal methods are used widely to validate these systems due to their high level of confidence for guaranteeing correctness, security, and safety properties. TLA+ is a formal specification language and its TLC model checker will be used in this work to formally verify and validate several example systems, and to find security hacks. Designing smart systems has recently received much attention because of the ability to connect heterogeneous sub-systems together to build a fully controlled smart system. Here we propose models for three systems: a smart school building system, an ADS-B system for avionics, and a SCADA system. We model the systems using UML diagrams to describe objects, states, and sequences of actions. Formal methods techniques are used to verify the properties of the systems. TLA+ is used for formal modeling. For the smart school building system, the ADS-B system, and the SCADA system, we verify their properties and correctness using the TLA+ toolbox and the TLC model checker. Also, we use the TLC model checker to find example security bugs we introduce into the designs.

Abstract: Aquaponics is a farming method that promises to be a good alternative against the food and environmental problem the world is facing. It is a combination between aquaculture (farming of fish) and hydroponics (growing plants without soil), being a technique to grow plants with the aquaculture effluent. This technique claims to have a high water efficiency, is pesticide-free, and reduces the use of fertilizers. All in all this technology is considered green and sustainable. Since the interest in aquaponics is increasing, the major challenge is to provide feasible and reliable solutions at the commercial scale. The concept of precision farming, usually applied in the traditional farming sense, is now being introduced, leading to the need to adopt sensing, smart, and IoT systems for monitoring and control of its automated processes. This thesis aims to support research towards a viable commercial aquaponics solution by first; identifying, listing, and providing an in-depth explanation of each of the parameters sensed in aquaponics, and the smart systems and IoT technologies in the reviewed literature. Secondly, to propose a tool that uses image-processing techniques, deep learning, and regression analysis to estimate the size of the crops as they grow using image segmentation and do a correlation between the size of the crops and their fresh weight for being modelled that will work as a performance metrics. Third, the development of a framework is presented that involves the creation of a wireless sensing module that uses sensing parameters and the connection to a database capable of storing and linking the information to a quality assessment tool. Finally, an application that adopts digital twinning in the growing beds of an aquaponics system for monitoring, in real time, parameters and hence control the aquaponics physical system is developed.

The Internet of Things (IoT) is, simply put, the concept of connecting any network enabled device to the internet and/or to each other. These IoT connected devices include household appliances, security cameras, health care monitoring systems etc. As IoT devices become smaller and more complex, identifying possible security threats has become a more resource intensive process. However, IoT devices have a limited amount of on-board resource capacity which must be utilized to perform ever-growing complex functions of the device. In other words, as functionally goes up security goes down. This security concern is amplified by the fact that IoT devices gather more and more sensitive data in regard to their users. Historically, device security has involved complex software systems which require massive amounts of system resource to run as well as a constant series of updates to identify new threats. This historical approach is not a viable solution for resource constrained data driven IoT devices. It is the goal of this study to: (a) identify IoT device sensory inputs with specific characteristics of interest, (b) introduce a novel approach for enhancing an IoT device’s security by building unique device-specific profiles based on sensory inputs, and (c) examine the potential use of such a profile for threat detection in IoT devices. To accomplish all three prongs of the goal a needed framework wasimplemented, and a viable authentication protocol was established. The research concluded by identifying two sensory inputs (power consumption and execution time) which possess the specific characteristics needed for device profiling. The identified sensory inputs were then used to create an IoT device-specific profile which can be used to enhance device security.