Digital optical fibre aided virtual MIMO systems in multicell multiuser networks

Abstract

Advanced communication networks are conceived by amalgamating the techniques of distributed antennas (DAS) and fractional frequency reuse (FFR), where the cell-edge areaadditionally relies on FFR for the sake of reducing the co-channel interference (CCI). Since DASs rely on the availability of a backhaul, we consider a realistic - rather than perfect- optical fibre backhaul and develop a model for a composite optical fibre and wireless channel, which is capable of representing the effects of the fibre induced imperfections imposed on the attainable throughput and the BER performance of the entire system. More explicitly, we demonstrate the effects of the fibre’s linear dispersion and nonlinearity on both noncooperative and cooperative DAS aided FFR schemes.

The proposed pervasive DAS/FFR scheme has a low complexity, since it employs a single omni-directional transmit element at each remote antenna (RA) and a single receive element at each mobile station (MS). The resultant system is jointly modelled as a virtual multiple-input and multi-output (MIMO). Hence, jointly designing the transmit pre-processing (TPP) matrix of all the cooperative RAs in the case of the downlink (DL) is shown to be beneficial. Similarly, the joint design of the multiuser detector (MUD) coefficients for all the active MSs transmitting in the uplink (UL) is carried out by the central processor of the base station (BS).

The dominant interference of our DAS/FFR scheme is caused by the intra-cell interference (ICI) for both cases of DL and UL, especially in the ’worst-case direction’ when the MS is roaming near the angle halfway between two adjacent RAs. In order to mitigate the ICI, the TPP matrix is designed for all the cooperative RAs is capable of achieving an increased throughput for the entire cell-edge area in DL, regardless of the specific geographic distribution of the users. Our novel combined probabilistic data association (PDA) multiuser detector is invoked by the mobile relays (MR) aided pervasive DAS/FFR architecture, which is capable of substantially reducing the bit-error ratio (BER) for the MS roaming at arbitrary positions, especially in the ’worst-case direction’. Practically, the generation of perfect Channel State Information (CSI) remains an open challenge, whilst having an imperfect CSI leads to a reduced performance. Hence, we also investigate the impact of practical impairments including the effects of CSI estimation errors, CSI quantisation errors as well as Orthogonal Frequency Division Multiplexing (OFDM) signal timing and frequency synchronisation errors in DL scenario. We further extended the Single-Input Single-Output (SISO) based non-coherent 16StQAM scheme to a Single-Input Multiple-Output (SIMO) system in UL scenario, when both the Turbo-Coded 16StQAM and 16QAM schemes in the multicell, multiuser uplink system considered without the perfect CSI. Finally, we consider a practical multiuser, multicell scenario, where a particular user’s position is mapped to a specific Signal-to-Interference-plus-Noise-Ratio (SINR). This method allows us to portray the geographic of the SINR across the entire cellular area. In the DL of the cooperative DAS aided FFR scheme, a throughput of h = 5bits/s/Hz may be maintained for an imperfect optical fibre backhaul, regardless of the specific geographic distribution of the users roaming in the cell edge area. Provided that an idle MR may be activated in the vicinity of the optimum relay position, in the UL of the cooperative DAS aided FFR scheme, 80% of the cell-edge area exhibits a BER, which is better than 10?4, while the remaining 20% has a BER value of [10^-4 · · · 10^-2]. Naturally, this BER performance improvement is achieved at the cost of potentially halving the throughput, because the MR has to receive and retransmit its information in different time slots. When applying power control (PC), the BER recorded across the entire cellular area may be reduced, below 10^-3 even without the assistance of MRs.

More information

Identifiers

ORCID for L. Hanzo: orcid.org/0000-0002-2636-5214

Catalogue record

Export record