Software Engineering for Internet of Things

点击翻译按钮获取中文摘要

The Internet of Things (IoT) is at the forefront of the new generation of software intensive technologies that are bound to deeply impact society. IoT applications interconnect a variety of sensors and actuators attached to people, buildings, cities, etc., and collect and process data in the cloud to improve people's lives, increase economic efficiency and ensure better environmental protection. Specific subdomains, such as autonomous vehicles and robots, smart transportation and buildings (generically known as Cyber Physical Systems), augment systems with actuating capabilities to change their behavior, structure or parameters. ***The long term goal of this research program is to develop scientific and engineering foundations for IoT systems that are predictable and robust. To achieve predictable and runtime manageable IoT systems, we need methods and tools to specify their runtime variability and constraints. The new generation of IoT systems is data intensive, where data is generated in large volumes and at high rates by dispersed sensors located at the edges of the infrastructure. It is costly and induces latency and variability to transfer the entire data to the public cloud. We focus on how to specify the data flows and adaptation components, such as sensors, actuators, analyzers and planners, so that they opportunistically place computation and transfer data, either in a decentralized or centralized manner. IoT systems are inherently adaptive at runtime: they monitor their environment, analyze the data and adapt to changes. The decision to adapt can be taken by the system itself when this is feasible or by a human-in-the-loop when the decision requires human intervention. Concurrently, there is the adaptive software system management, where software components are automatically reconfigured, optimized or protected or repaired at runtime. We will explore the design, implementation, interference and eventual coordination of these adaptive services so the overall system requirements are met. Fundamental to building robust IoT applications are models of the software and of the physical world. At design time, models can allow the designer to explore the behavioral space and verify if specifications are met. At runtime, models will serve runtime adaptive controllers to manage both the software and the physical parts. ***This research program will use theoretical and experimental methodologies and train students in the emergent fields of Internet of Things, adaptive systems and cloud computing. The training will prepare them for jobs that are in high demand in the Canadian economy. The program will also equip the students with the knowledge to understand the social and ethical implications of the new technologies. It is expected that the results of this research program will advance the state of the art in areas that are of great importance to Canadian industry, researchers, and citizens.**