Factor graph-based receivers for multi-carrier transmission in two-way relaying and massive machine type communications

Abstract

The way of communication has evolved in recent years from purely human-driven communication - with the need for voice calls and simple data packets such as SMS - to a mixture of fast human-driven communication and machine-driven communication, leading to networks that support a growing number of services with different quality of service requirements and an increasing number of devices. This development is at the expense of a wide range of different service needs that have serious implications for the architecture of mobile communications systems. Some forecasts predict an explosion of connected devices in the near future. This extremely high demands on today's communication systems are constantly increasing as very broad requirements such as coverage, short latency times and simple transmitter design are made. In particular, the latter leads to cheap and therefore imprecise hardware implementations, leading to higher detection processing at a receiver. The aim of this thesis is to analyze such scenarios using a framework of robust multicarrier transmission systems and factor graphics-based detection schemes. Overall, this work gives a deeper insight to multi-carrier schemes applying well-localized waveforms and its applications in Two Way Relaying and massive machine type communication. Furthermore, for massive machine type communication a multi-carrier narrow band system that addresses key aspects in this type of communication, which takes advantage of compressive sensing multi-user detection has been introduced. Especially, for these scenarios the framework of well-localized waveforms and factor graphs, utilizing the knowledge of the waveforms, lead to flexible and controllable design of factor graph-based equalizers. The results of these analyses yield the statement that multi-carrier concepts with well-localized waveform can improve the robustness in multiple access schemes, as shown by performance investigations with time and frequency offsets in future mobile radio systems.