Performance Evaluation of Device and Machine-Type Traffic in Wireless Networks
Amirshahram Hematian
Dissertation Committee
Wei Yu (Chair), Chao Lu, Michael McGuire Alexander Wijesinha
Time: 1pm, July 25th, 2018
Location: Computer and Information Science Dept. Conference Room (YR 459)
A modern wireless network provides ubiquitous networking connections to devices and services to support diverse critical applications. These applications demand massive data traffic transmitted from a large number of devices. However, there are limitations to expanding data service and traffic capacity to support all types of data traffic, such as Internet-of-Things (IoT), video, voice and others, by only boosting the system capacity via adding more antennas, sub-carriers, symbols, slots, etc. To address these challenges, in this dissertation we first develop a clustering algorithm to leverage Wi-Fi Direct (as an outband solution) and enable Device-to-Device (D2D) communication, which can not only offload massive data traffic from the LTE (Long Term Evolution)-based cellular network, but can also support the communications of Internet-of-Things (IoT) applications. Second, we create a Software Defined Radio (SDR) testbed and conduct real-world LTE network case studies, including optimizing video transmission via adaptive LTE links to minimize over-allocation of LTE links, performing User Equipment (UE) location estimation by leveraging the signal strength from nearby cell towers where GPS is not available, and others. Third, to deliver large volumes of IoT data with high data rates generated by Smart Meters (SMs), we design a Hardware-In-the-Loop (HIL) simulation environment that can interact with a real-world cellular network to assess the effectiveness of transmitting the SM data over LTE.
