Our research covers the themes of Smart Transportation, Smart Buildings, Remote Assisted Living, End-to-End IoT Analytics, DevOps and Security, Privacy and Ethics for IoT.
Traffic congestion reduces the productivity of urban regions, increases costs through additional fuel consumption and carbon footprint and contributes to air pollution. IoT is instrumental in Integrated Corridor Management (ICM), which can improve urban mobility and safety. This theme will investigate the design of near-term ICM systems from the applications and infrastructure perspective. The focus of the research will be in (1) the unification of a variety of IoT sensing devices providing audio, image, video, air quality, and weather data streams, together with mobility information from trains and public transit, and other mobility detectors for cars, bicycles, pedestrians and crowds; and (2) the design of distributed processing of these data streams combining fog, edge, and core processing to realize ICM. The research will combine algorithm and architecture design with experimental testing and evaluation using deployments in partner cities in the Greater Toronto Area.
Smart buildings integrate a level of monitoring, system integration, and system automation that exceed typical levels from the early 2000s with the goal to improve operational effectiveness and user comfort while decreasing energy and associated carbon emissions. The scope of this research theme includes the full stack of IoT-related technologies required to obtain (sensor integration), ingest and stream (data acquisition architecture), store and retrieve (information schema and query development), visualize (using BIM and other 4D/nD tools), predict (machine learning as well as advanced energy and process simulation tools), evaluate and control building performance, security and privacy.
Remote Assisted Living
Canadians living with chronic and debilitating conditions face many challenges that limit their function and independence. In this project, we will develop technologies and services to support people in maintaining their functional independence in spite of these conditions. One key inhibitor of such ventures is the existence of a plethora of devices with various standards and protocols in connectivity and data transfer, which makes their integration challenging. Our focus is on developing an extendible and adaptable smart-home platform, on which to deploy and integrate such IoT devices. This platform will consolidate and analyze the data emitted from these devices to enable activity recognition, exceptional event identification, risk assessment and alert generation, activity recommendation and prompting, and automated device control. The complexity of the infrastructure underlying these smart homes demands advanced human-machine interaction models. Relying on model-driven engineering methodologies, we will develop high-level end-user programming languages, for home-owners, occupants, and caregivers to configure the platform and manage its services.
End-to-End IoT Analytics and Assurance
IoT systems emit Big Data with two consequent challenges. The first challenge involves quality assurance for functional and non-functional properties (i.e., operational qualities such as availability, efficiency, performance, reliability, robustness, security, stability, and usability). This is due to the variety and veracity of monitoring data (hardware, network, software), as well as the volume and velocity of reaction required for real-time adaptation. The application domains we focus on are natively characterized by large amounts of data, where high-speed analytics are also crucial. To give an example, in domains like Assisted Living or Autonomic Driving, real-time analytics in terms of seconds can be the difference between life and death. By end-to-end analytics, we mean the focus of this research in providing fast, reliable and robust analytics from the development and management of the IoT system and its infrastructure, all the way to the application domain itself. The premise of end-to-end IoT analytics is to identify patterns and exceptions in heterogeneous data sources that will allow us to advance scholarship, improve the human condition, and create commercial and social value.
DevOps for IoT
IoT applications rely on continuously evolving objects of great heterogeneity and variability. For instance, the traffic information acquired from a transportation system relies on an ever-changing ecosystem of smart objects, including sensors, cars, trucks, traffic lights, belonging to different vendors or stakeholders, adhering to different standards and having changing capabilities. Developing, deploying and maintaining IoT systems—involving hardware devices, heterogeneous environments, numerous software components, network infrastructure, feedback loops and dynamic big data—is enormously challenging. Traditional software engineering and operation management approaches are not designed with such infrastructure complexity, heterogeneity and fluidity in mind. DevOps is a promising set of software engineering processes, methods, techniques and tools to systematically orchestrate highly dynamic software systems, such as IoT systems. The goal of DevOps is to combine the processes of continuous development and operations management of already deployed systems. By leveraging models at runtime (MART) and control theory, DevOps can equip IoT systems with capabilities for autonomous management and runtime adaptability including performance and security objectives. This is crucial for systems that operate within highly dynamic and high risk environments. In addition, Blockchain networks will be used to manage trustworthy and secure data flow within a heterogeneous ecosystem with various stakeholders.
Security, Safety, Privacy and Ethics in IoT
Intertwined with the great potential of IoT are a number of inherent risks that could arise if the integration of IoT devices into the wider Internet and Cloud infrastructure is not approached cautiously with regard to the issues of security, safety, privacy and ethics (SSPE). SSPE are key pillars of any technology that aims to be effective and dependable. Consequently, failing to address SSPE adequately could jeopardize the successful adoption of IoT. Protection of IoT devices will protect the users and the infrastructure (e.g., by hacking a particular device of an IoT ecosystem, the attacker could steal all of its internally stored information, thus compromising the privacy of its users). Moreover, adequate protection of IoT devices is also of great importance for the Internet and Cloud infrastructure at large; for example, with Mirai botnet (2016) unprotected IoT devices were used to launch extremely powerful attacks against various targets throughout the Internet [20,21]. In this theme, we will conduct SSPE research with two objectives. (1) to understand and analyze various threats/attacks specifically arising from a number of unique limitations pertinent to the design and operation of IoT devices; (2) to design, develop and validate effective protection mechanisms against the above threats/attacks using security risk management and machine learning (so these solutions are reached in an autonomous manner).