Receiver Architecture Analysis and Design for OFDM Based Communication Systems

  • Fatma Kalbat

Student thesis: Doctoral Thesis

Abstract

Orthogonal frequency division multiplexing (OFDM) is the dominant technology in the fourth generation (4G) long term evolution (LTE) networks. With an appropriate cyclic prefix (CP), OFDM is able to combat the delay spread of wireless channels and eliminate inter symbol interference (ISI) and inter carrier interference (ICI). OFDM has several advantages such as low complexity equalization, easy adoption to multiple input multiple output (MIMO) transmission, simple channel estimation and improved spectral efficiency in frequency-selective multipath fading environments. Although OFDM has a series of attractive properties and features, it is not exempt of defects and drawbacks such as sensitivity to carrier frequency offset (CFO), timing offset (TO) and high peak-to-average-power ratio (PAPR). The adoption of traditional OFDM in the next generation of wireless networks, also denoted as the fifth generation (5G), is not taken for granted, where it is unable to meet many new demands required for 5G networks. The deployment of wireless networks has increased significantly with the widespread use of smart phones and changed qualitatively in association with Internet of Things (IoT). Therefore, 5G is expected to satisfy the requirements of a high data throughput, low latency, low cost, and increased spectral efficiency. To address the new challenges in 5G networks, it is necessary to study the competence of existing OFDM systems and seek possible enhancements. In this context, this dissertation focuses on the analysis of OFDM based systems in different practical scenarios and propose new solutions for some of the challenges. More specifically, we consider linearly precoded OFDM (P-OFDM) systems to resolve some of the main limitations that face conventional OFDM, and evaluate the performance of P-OFDMin the presence of certain system imperfections. In the first part of this thesis, we present a new MIMO technique, based on Walsh Hadamard Transform (WHT) precoding, to improve the robustness of space-frequency block coded OFDM (SFBC-OFDM) systems. The WHT is applied to the data symbols prior the Alamouti encoder at the transmitter and to the output of the Alamouti decoder at the receiver. The computational complexity of the proposed system is evaluated in terms of complex additions and multiplications where the numerical results show that the proposed system has a lower complexity as compared to other P-OFDM systems. Moreover, the proposed system is highly robust to the channel time and frequency selectivity as compared to conventional SFBC, space time block coded (STBC) and other P-OFDM systems. The second part of the thesis considers the bit loading problem for communication systems that utilize OFDM in conjunction with precoding (P-OFDM) or time-domain interleaving (I-OFDM). In particular, we propose a new bit loading algorithm for P/I-OFDM that has substantially higher effective throughput and less computational complexity, when compared to bit loading in conventional OFDM systems. The obtained results show that the effective throughput of P/I-OFDM can be more than fourfold the conventional OFDM while the complexity is less than 1.5%. Moreover, the results show that the PAPR properties of the considered systems are preserved under the adaptation process. The third part of the thesis provides the performance evaluation of P-OFDM in the presence of CFO. The performance of the considered P-OFDM is evaluated in terms of the signal to interference plus noise ratio (SINR) and bit error rate (BER). Various channel models are considered and closed-form analytical expressions are derived for the exact SINR. Closed-form analytical expressions are derived for the exact SINR in AWGN and flat fading channels and semi-analytical SINR expression is provided for frequency-selective channel case. The obtained analytical results, corroborated by simulation, show that P-OFDM is substantially more sensitive to CFO compared to conventional OFDM. Generally speaking, if the normalized CFO is about 18%, OFDM will outperform P-OFDM for most practical channel models. It is also interesting to note that the subcarriers in P-OFDM may experience different SINRs, which is not the case for conventional OFDM. This work also considers the SINR of OFDM after the equalization process. The obtained results show that ignoring the impact of the equalization process results in inflated SINR degradation. Finally, the fourth part of the thesis considers the performance evaluation of P-OFDM in the presence of timing offset (TO). WHT is used to perform the precoding process, and the performance is evaluated in terms of SINR and BER. Closed-form analytical expressions are derived for the exact SINR in additive white Gaussian noise (AWGN) and flat fading channels, while the SINR is obtained using Monte Carlo simulation in frequency-selective fading channels. Moreover, Monte Carlo simulation is used to corroborate the derived analytical expressions. The obtained results show that P-OFDM have substantially different performance in the presence of TO as compared to conventional OFDM where the SINR is different for each subcarrier. Consequently, subcarriers with severe SINR degradation will experience a drastic BER increase, which may drive the overall BER to become worse than conventional OFDM. Additionally, it is interesting to note that P-OFDM is more sensitive to positive TO than negative TO in frequency-selective channels. The SINR performance of equalized OFDM is considered in this part.
Date of AwardJun 2019
Original languageAmerican English

Keywords

  • OFDM
  • MIMO
  • precoding
  • bit loading
  • synchronization.

Cite this

'