M2M services, as the key enabler of smart city business, have traditionally been delivered in domain-specific and vendor-locked platforms. Although this model has successfully supported the growth of M2M business to date, it is facing strong challenges in terms of elasticity, device integration, and application development. PC3L is developing a prototype of cloud-based M2M service delivery platform. The platform will not only seamlessly integrate with cloud infrastructure to achieve elasticity, but also open itself in PaaS paradigm to foster new business models in the convergence of M2M and cloud computing. It will support vertical M2M solution development and end-to-end service delivery, as well as self-served open M2M application development.
Pervasive computing environments have evolved from hardware centric to data and application centric. Ubiquitous sensors are generating enormous amount of data, and a broad spectrum of applications is dependent on the data acquired from the environment. The emergence of data intensive pervasive applications, such as traffic scheduling and building energy management, is posing unprecedented research challenges—-to acquire and process large amount of spontaneous data with QoS assurance. We are focusing on three key aspects of data processing: 1, programming model of data processing plans in large-scale pervasive systems; 2, self-adaptive and scalable sensory data processing middleware; 3, data quality assurance and optimization in volatile environments.
Sensory data contains valuable information. For example, in smart building automation, the energy consumption patterns are critical for sustainable management and cost savings. However, sensory data come at large amount with unreliable quality, making them unsuitable for knowledge discovery tasks. This research is intended to create methods dedicated to knowledge modeling and discovery on sensory data. Furthermore, efficient algorithm execution approaches on data processing middleware is also under investigation along with the development of data processing middleware. We are exploring several data mining approaches on real-world sensory datasets for quantifying the sustainability of smart buildings.
Sustainability governance is a complex subject that requires being able to capture, integrate, and maintain relevant data from different domains, covering economic, social, and environmental aspects, over time in order to enable the analysis and compliance monitoring of complex sustainability measurements. Furthermore, the analysis and monitoring also requires the development of novel algorithms to analyze, simulate and predict the future patterns of these complex sustainability measurements by utilizing multi-domain data and related computational models. The PC3L focuses on fundamental techniques and frameworks for management, integration, composition and utilization of data, algorithms, services, and compliance processes in large-scale environments.