The Analysis And Compensation Of Phase Noise In The Vector OFDM Wireless Air Interfaces

Phase noise is highly detrimental to the performance of high speed communication systems as it results in random rotations of the signal constellation from symbol to symbol. Phase noise as one of the RF analog front end non-idealities, has its origins from the local oscillators (LO) that are used in the UP and DOWN conversion process that is present in all wireless transceivers. Emerging wireless communication systems employ higher bandwidths, higher constellation orders and mul-tiple antennas to meet the increasing market demands. This further imposes stringent requirements on the local oscillators and thus driving up the cost of the analogy front end. In-order to minimize this cost, it is desirable to be able to tolerant a certain amount of the phase noise ideality in the front end and then estimate and compensate for it in the digital domain at a reasonable complex-ity. In order to meet new user demands, the past decade has seen the emergence of new physical (PHY) layer wireless air interfaces or transmission standards such as Orthogonal Frequency Divi-sion Multiplexing (OFDM), Single Carrier Frequency Domain Equalization(SC-FDE), Multiple-Input Multiple-Output(MIMO), Vector Orthogonal Frequency Division Multiplexing (V-OFDM), Space Time Block Coding (STBC) and many others. Several publications have addressed the per-formance of each of these standards in the presence of phase noise. However, research is already in progress for possible fifth generation (5G) PHY layer wireless air interfaces to be employed in the5G systems because of the limitations that3G and4G based PHY layer wireless air inter-faces impose. Orthogonal Frequency Division Multiplexing (OFDM) has almost become a defacto standard in most of the3G and4G wireless interfaces mainly due to its robustness against multi-path channels and easy implementation based on the fast Fourier transform (FFT). Single Carrier Frequency Division Multiplexing SC-FDE has also been adopted in some standards to overcome some of the short comings of OFDM such as a high Peak Amplifier Power Ratio (PAPR) and the reduced performance of OFDM when high constellation orders are employed. The definition of two PHY layer interfaces makes the system design to be rather complicated and asymmetrical. As a bridge between OFDM and SC-FDE, Vector Orthogonal Frequency Division Multiplexing (V-OFDM) has been shown as one of the most possible candidates for the up-coming5G wireless air interfaces mainly because of the flexibility and superior performance that it offers. Generalized Orthogonal Frequency Division Multiplexing G-FDM which is based upon (V-OFDM) is also a subject of major research. Whilst the phase noise impairment has been greatly studied in the3G and4G based wireless air interfaces, little or no research has been done to study how radio fre-quency (RF) impairments might affect the performance of the upcoming5G interface standards. It is the aim of this thesis to study the effects of the phase noise impairment in V-OFDM systems which is a promising candidate for future generation wireless air interfaces. Furthermore, other promising5G interfaces such as GFDM are based on V-OFDM. In OFDM systems, phase noise leads to a Common Phase Error (CPE) as well as an Inter Carrier Interference (ICI) effect whilst in SC-FDE phase noise results in random rotations of the transmitted symbols. In this thesis a mathematical model that fully describes the effects of the phase noise in V-OFDM systems is de-rived. We show that phase noise in V-OFDM systems leads to a Common Vector Block Phase Error (CVBPE) as well as an Inter Vector Block Carrier Interference (IVBCI) effect. A closed-form general expression for the Signal to Interference Noise Ratio (SINR) in V-OFDM systems is derived. SINR analysis shows that V-OFDM systems SINR behaviour is pretty similar to that of OFDM systems. However, V-OFDM always has higher SINR floors as the phase noise impairment is increased which is highly desirable. Novel pilot based phase noise estimation and compensation techniques are developed. A complexity analysis of the techniques shows that V-OFDM systems have an extremely low complexity as compared to OFDM systems. Furthermore, after the phase noise compensation, V-OFDM systems have superior performance as compared to OFDM sys-tems. Blind and semi-blind phase noise estimation schemes that use very few pilot symbols or no pilot symbols at all are developed. The two schemes are also shown to perform exceptionally well. The analysis is validated through the use of Monte Carlo simulations.