Use this URL to cite or link to this record in EThOS:
Title: Compressed sensing-aided multi-dimensional index modulation transceiver design
Author: Lu, Siyao
ISNI:       0000 0004 8510 3833
Awarding Body: University of Southampton
Current Institution: University of Southampton
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
In this thesis, we propose a suite of compressed sensing (CS)-aided index modulation (IM) techniques for transmission over frequency-selective fading channels invoked in orthogonal frequency division multiplexing (OFDM) and multiple-input multiple-output (MIMO) arrangements. The design objective is to improve both the performance and the design flexibility by striking flexible trade-offs among the performance, bandwidth efficiency, energy efficiency and complexity. Specifically, a novel CS-aided frequency-domain IM system is proposed, combining the benefits of space-time shift keying (STSK), OFDM-IM and CS, in order to strike a flexible trade-off between the throughput and bit error rate (BER) performance. The proposed CS-aided frequency-domain IM is termed as CS-OFDM-STSK-IM. Explicitly, the CS-aided signalling strategy is applied to the transmitter for improving the throughput by introducing the concept of virtual domain. Then a pair of reduced-complexity detectors are proposed, which rely on the CS principles. In particular, we derive the discrete-input continuous-output memoryless channel (DCMC) capacity for characterising the achievable performance limit of the CS-OFDM-STSK-IM. Furthermore, in order to attain extra diversity gains at no cost in terms of power or bandwidth, the coordinate interleaving technique is applied. Additionally, for the sake of achieving a near-capacity performance, we conceive a two-stage serially concatenated channel-coded CS-OFDMSTSK-IM scheme employing iterative detection. To attain an improved BER performance as well as system capacity in comparison to the CS-OFDM-STSK-IM scheme, we propose an improved CS-aided multi-dimensional IM system relying on both frequency- and spatial-domain IM. Explicitly, extra implicit information bits are transmitted by detecting the activated indices of both the TAs and subcarriers, while striking a flexible design trade-off between the throughput and the diversity gain. Furthermore, CS is invoked at both the transmitter and the receiver of our multi-dimensional system for the sake of improving the system's design flexibility, whilst reducing the detector's complexity. We first present the maximum likelihood (ML) detector of the CS-aided multi-dimensional IM system for characterising the lower bound of the system's performance. Specifically, an upper bound is derived for the average bit error probability (ABEP) and it is observed that the theoretical upper bound derived becomes tight when compared to the ML detector simulation curves as the value of SNR increases. Although the ML detector is capable of achieving the optimal performance, its prohibitive computational complexity makes it impractical for our multi-dimensional scheme. Hence we propose a reduced-complexity detector based on CS principles, which imposes only a modest BER degradation compared to the ML detector. We also analyse the computational complexities of both the ML and of the reduced-complexity detectors in different system configurations. We show that the CSaided reduced-complexity detector is capable of attaining a flexible trade-off between the throughput and the complexity. Furthermore, soft-input soft-output decoders relying on both the optimal ML and on the reduced-complexity detectors are proposed for attaining near-capacity performances, which are analysed with the aid of EXtrinsic Information Transfer (EXIT) charts. Then we characterise the maximum achievable rate of the CS-aided multi-dimensional IM using both the ML and our reduced-complexity detectors by formulating the DCMC capacity and by invoking EXIT charts. Finally, attracted by the benefits of large-scale multi-user MIMO (LS-MU-MIMO) systems in terms of their high bandwidth efficiency, high energy efficiency and increased reliability, we conceive a novel two-dimensional IM, namely the CS-aided (generalised) space-frequency IM (CS-(G)SFIM) scheme in our LS-MU-MIMO uplink scenarios. Explicitly, extra information bits are conveyed by mapping both to the activated spatial and frequency-domain indices. More particularly, we employ different transmit antenna (TA) activation patterns for different transmitted symbols instead of using the same active TAs for the whole subcarrier group as proposed in the CS-aided multi-dimensional IM scheme, which improves both the BER performance as well as the diversity gains. For the sake of further increasing the system's achievable rate without sacrificing any bandwidth or energy, CS-aided pre-processing is applied to the frequency-domain IM. Then we design a CS-aided reduced-complexity multi-user detector, namely the reduced search-space based iterative matching pursuit (RSS-IMP), which significantly reduces the detection complexity compared to the ML detector. Furthermore, the RSS-IMP detector is capable of significantly reducing the detection complexity, while attaining better performances than both the conventional MU-MIMO-OFDM system and the proposed system using the classical minimum mean square error (MMSE) detector. We also characterise the performances of the CS-(G)SFIM system in the presence of channel estimation errors. Our simulation results show that the CS-(G)SFIM system is more robust to imperfect channel estimation than the conventional MU-MIMO-OFDM system. More importantly, owing to the highly flexible structure of our proposed scheme, improved trade-offs may be struck among the bandwidth efficiency, energy efficiency, BER performance, as well as the computational complexity by beneficially configuring our design, depending on the associated system requirements and channel conditions.
Supervisor: El-Hajjar, Mohammed Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available