Joint OSC Receiver for Evolved GSM/EDGE Systems

Abstract

This paper is focused on evolved GSM/EDGE systems complying to the new feature Voice services over Adaptive Multi-user channels on One Slot (VAMOS), recently introduced in the 3GPP GSM/EDGE radio access network (GERAN) standard. VAMOS enables the transmission of two GSM voice streams on the same radio resource through the so called Orthogonal Sub Channel (OSC) multiple access technique which aims at doubling the number of users served by a cell. When adopting this feature, the GSM network experiences a mixture of in-cell and out-of-cell interference which has to be handled by advanced receivers with interference rejection capabilities. In this paper, a novel two-stage receiver is proposed and tested for the uplink of GSM VAMOS systems. The new scheme combines a pre-filtering stage for out-of-cell interference mitigation with a multi-user detector (MUD) for joint equalization of the two multiplexed OSC streams. The approach is suited for either single-antenna or multi-antenna base stations and can accommodate multiple filters for each user to cope with complex multipath propagation scenarios. The new solution is compared with other existing interference cancellation techniques here applied to the specific OSC scenario. Numerical results show significant performance gains in realistic multi-cell simulated scenarios

Enhanced Widely Linear Filtering to Make Quasi-Rectilinear Signals Almost Equivalent to Rectilinear Ones for SAIC/MAIC

“…The structure of the generic M -input pseudo-MLSE receiver (M = 1, 2) is then depicted at Figure 1. It is composed of the TI WL filter (20), which reduces to a linear filter for M = 1, followed by a sampling at the symbol rate, a real part capture (for M = 1) and a decision box implementing the Viterbi algorithm, since r *…”

“…Expression (35) then shows that for M = 2 and for both R and QR signals, the WL filter (20) performs SAIC for SOI and CCI propagation channels such that | α ISF (f )| is not constant and equal to 1 over B 0 F , i.e. for most of propagation channels.…”

“…When |α sI | = 1, i.e. when there exists a spatial discrimination between the SOI and the CCI (which requires N > 1), assuming a strong CCI (ε I 1), we obtain from (20), (28), (31), (32), (37), (38), and after straightforward derivations, the following expressions:…”

“…The separation, at the handset level, of the two streams, distorted by frequency selective channels and potentially corrupted by co-channel OSC and/or legacy GMSK interference, coming from out of cell OSC and/or legacy base-stations, requires the implementation of enhanced SAIC techniques for QR signals [16]. Such preliminary enhanced techniques for VAMOS, based on standard WL filtering, have been proposed recently in [17][18][19] for SAIC and in [20] for SAIC/MAIC, for both OSC downlink and uplink transmissions respectively.…”

“…In this context, the purpose of this paper is twofold. The first one is to get more insights into the existing SAIC/MAIC technology by proving analytically, which is completely original, that, contrary to what is implicitly accepted as true in the literature [6], [7], [9], [17][18][19][20], [29][30][31], [34][35][36], QR signals may be less efficient than R ones for SAIC/MAIC implemented from some standard WL filtering. Starting from this result, the second purpose of the paper is to propose and to analyse, partially analytically, which is also very original, for QR signals and frequency selective fading channels, an enhanced SAIC/MAIC technique based on a three-input WL FRESH receiver.…”