Use this URL to cite or link to this record in EThOS:
Title: Compressed sensing aided design aspects of indoor attocells
Author: Zhang, Hongming
ISNI:       0000 0004 7225 136X
Awarding Body: University of Southampton
Current Institution: University of Southampton
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Compressed Sensing (CS) is invoked for indoor attocells utilising Power Line Communication (PLC) and Visible Light Communication (VLC), in order to strike a compelling tradeoff between the performance attained and the complexity imposed. We commence by detailing our motivation as well as by introducing PLCs and VLCs. Then, we specify the related technical issues to be addressed and outline the general framework of our CS-assisted system design. Firstly, we propose a Compressed Impairment Sensing (CIS)-assisted Interleaved-Double-FFT (IDFFT) system for broadband PLC, in order to mitigate the deleterious effects of asynchronous impulsive noise. A Frequency-Domain (FD) equaliser is employed for achieving a beneficial multipath diversity gain, which is amalgamated with a Two-Dimensional (2D) interleaver conceived for effciently dispersing the asynchronous impulsive bursts encountered. As a result, the attainable system performance of the IDFFT system becomes better than that of the classic Orthogonal Frequency-Division Multiplexing (OFDM) system. As a further improvement, some of the disabled sub-bands, which are due to either high level of attenuation or narrowband interference, are exploited as pilot symbols in our IDFFT system for supporting the operation of CIS. We demonstrate that the sparse impairment vector can be readily estimated by our 2D-interleaving aided CS solution. Moreover, we propose a computationally efficient search algorithm capable of approaching the performance of the optimal pilot design. Finally, based on the turbo principle, we propose a joint CIS estimation and symbol detection algorithm for our IDFFT system. Explicitly, the CIS estimator iteratively exchanges ever more reliable information with the symbol detector upon each new iteration, hence resulting in an improved system performance. More particularly, our CIS estimator invokes a greedy algorithm, where locally optimal choices are made at each iteration. The performance of our CIS-assisted IDFFT system is investigated by simulations, where the various aspects of the design-tradeoffs are studied. We demonstrate that our CIS-assisted detector is capable of efficiently mitigating the deleterious effects of asynchronous impulsive noise at a low complexity. Secondly, we propose a Linear Precoding assisted Index Modulation (LPIM) scheme and a Compressed Sensing assisted Index Modulation (CSIM) scheme for benecial bit-to-symbol mapping in OFDM systems. These mapping schemes are represented by a pair of codebooks. We commence by analysing the system performance in the context of achievable Spectral Efficiency (SE), Energy Efficiency (EE), diversity gain and coding gain. Moreover, the detailed codebook design criteria are outlined and discussed. Then, our LPIM codebook is designed based on the maximum diversity gain and the maximum coding gain criteria. By contrast, our CSIM codebook is designed based on the maximum diversity gain and the maximum achievable rate criteria. Both analytical and simulation results are provided for characterising the attainable performance of LPIM and CSIM, demonstrating that LPIM is capable of attaining an attractive coding gain, whilst CSIM is capable of striking a compelling tradeoff between the attainable coding gain and the maximum achievable rate. Additionally, we also demonstrate that in comparison to the conventional Subcarrier Index Modulation (SIM) scheme, our CSIM is capable of achieving a higher SE as well as an increased EE. However, the complexity of Maximum-Likelihood (ML) detection of both the LPIM and the CSIM symbols may be excessive, hence we propose a low-complexity Generalised Iterative Residual Check Detector (GIRCD). Similar to the CIS estimator, our GIRCD is also based on the basic principle of greedy algorithms. Finally, simulation results are provided for demonstrating that the GIRCD is capable of striking a beneficial tradeoff between the Bit Error Ratio (BER) performance and complexity. Finally, our CSIM scheme is applied to Direct Current biased Optical OFDM (DCO-OFDM) for VLC relying on Intensity-Modulation and Direct Detection (IM/DD), in order to improve the maximum achievable rate of the optical system. Both analytical and simulation results are provided for characterising the attainable performance of the CSIM-assisted DCO-OFDM system, showing that for the same maximum achievable rate as well as the same sensitivity to nonlinear distortions, the BER performance of the DCO-OFDM using CSIM is better than that of the DCO-OFDM using Quadrature Amplitude Modulation (QAM). In particular, we demonstrate that about 24% improvement of the maximum achievable rate expressed in bits-per-second-per-Hertz can be achieved by our CSIM scheme using Quadrature Phase-Shift Keying (QPSK), in comparison to the conventional QPSK scheme dispensing with our CSIM arrangement. As a further improvement, a piecewise Companding Transform (CT) is conceived for our O-OFDM system, forming the CTO-OFDM arrangement for the sake of mitigating the nonlinear distortions induced by the limited linear dynamic range of Light Emitting Diodes (LEDs). We rst investigate the general principles and design criteria of the piecewise CTO-OFDM scheme and conceive three prototype designs. Furthermore, we investigate the nonlinear effects of both hard clipping and CT on our O-OFDM systems by both analytical and simulation results. Our investigations show that CTO-OFDM constitutes a promising signalling scheme for VLCs, as a benefit of its high bandwidth efficiency, high reliability, high implementational flexibility, as well as robustness to nonlinear distortions.
Supervisor: Yang, Lieliang Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available