Abstract
The focus of this dissertation is the new emerging domain of Internet-of-Things (IoT) where devices, ranging from smart watches and smart textile to smart cars, are interconnected through several wireless/wired networks. In the specific research domain there are three main research pillars:
• Devices including sensors and actuators
• Communication based on wired/wireless networks and
• Efficient data processing and decision support
Device heterogeneity is one of the major problems in IoT world, which numbers billions, or even trillions of devices which are expected to become interconnected worldwide over the next years. And this is the first goal of the current dissertation which intents to provide a framework for allowing heterogeneous sensors to become interoperable in a seamless and efficient way.
The second challenge that will be addressed is the communication among the heterogeneous devices, which usually relies in a combination of short-range communications like Bluetooth and wired/wireless networks. Given that the devices are mobile, both indoors and outdoors, there are several communication challenges like error prone communication challenge, dynamic and unpredictable topology changes, diverse QoS support, diverse security support and of course diverse throughput support, which must be addressed. As part of this dissertation research directions like time constrained performance, enhanced communication robustness, minimization of power wastage, end-to-end link admission control and adequate gateway design towards common end to end QoS notion will be explored in order to propose an efficient communication mechanism allowing QoS over the proposed IoT framework.
The third research area that will be addressed is the efficient processing of the data collected for supporting decision-making. Data processing can be either at the device level or at the level of the IoT infrastructure (e.g. fog computing, cloud computing). As part of the proposed IoT framework, innovative techniques will be devised and proposed that will allow the efficient (in terms of power, speed, communication and QoS) data processing.
The target areas of the proposed IoT framework are demanding application domains including medical applications, robotics and smart home environments.
Subject of study:
This study has a main objective to review and analyze all relative communication technologies, platforms, OSs, and programming environments. Firstly, we focus on the evaluation of interoperability challenges concerning specific technologies as well as the simulation based or real platform experimentation. We propose new architectures, which will be designed, implemented and evaluated, that can offer homogeneity among different and diverse technologies. We will also propose approaches (which will be designed from the beginning, implemented and evaluated) that yield high efficient processing capabilities, while taking into consideration scarce resources, time constrained requirements and semantic correlation.
Advisory Committee:
Supervisor: Theofanis Orphanoudakis, Associate Professor, HOU
Nikolaos Voros, Associate Professor, University of Peloponnese
Achilles Kameas, Professor, HOU