Co-channel interference and inter-symbol interference (ISI) are common problems encountered by modern wireless communication systems. To mitigate these challenges, diversity antennas with optimum combining can be used in conjunction with the maximum likelihood sequence estimator (MLSE) to achieve the optimum receiver in a time dispersive propagation environment. We derive a simple multi- antenna MLSE receiver structure that is optimum in diversity combining and equalization given the spatial correlation of the interference and noise. The Viterbi sequence estimation within the multi- antenna receiver has the same complexity as a single-antenna receiver. Since the spatial covariance of the noise is taken into account in the ML sequence estimate, strong co-channel interference can be effectively suppressed. The receiver's robust interference rejection capability is validated in simulations of the GMSK-modulated GSM signal.

The paper considers the impact of the UMTS synchronisation channel (SCH) on bearer performance. The SCH is described, and it is shown that it has significantly different characteristics from other interference and noise sources. For this reason, it is important to understand whether it can be generally modelled in a simple way, e.g. by using its average received power. In reality, it is found that this is not the case, and that the performance degrades when the SCH is present for the same average level of E/sub b//N/sub t/. This shows that the coding/interleaving schemes used are not sufficient to average the temporally selective interference. In particular, it is found that convolutionally coded bearers show higher sensitivity than turbo coded ones, resulting in an increase in the E/sub b//N/sub t/ targets set by the outer loop. A basic analysis indicates that the resulting system capacity loss is not significant. However, a full evaluation of this effect is going to require a more complex analysis based on system simulation.

This paper presents results of an adaptive antenna measurement platform for DCS1800 frequency division duplex operation. A description of the measurement site in central London, UK is followed by a brief note on the data available and used for post-processing in this paper. Frequency division duplex operation is analysed in terms of the estimated peak-energy spectral energy density 'look-direction' in the uplink and downlink channels and compared to GPS estimated positions of a mobile transmitter. Singular point measurements from two specific locations in central London are analysed in terms of delay, Doppler and angular spread. It has been found that the speed of the mobile to duration of the measurement period ratio determine not only the fading statistics but the correlation in the uplink and downlink AOA estimates. Statistical analysis of the angular accuracy of the uplink direction estimate to the GPS estimated angular estimate shows a standard deviation of 35 degrees. It is possible to conclude that DOA based location estimators in a highly scattered environment and using four antenna element panels will be inaccurate to such an extent that DOA based location estimation is probably not viable for E911 requirements in an European context. This location estimate at the 67% percentile is within an angle error, with respect to the 'true' GPS estimated angle of arrival, of /spl plusmn/12 degrees for high mobile speed/measurement period duration ratios and /spl plusmn/30 degrees for all mobile speeds.

Radio-fibre microcellular networks are composed by small-size cells that make use of a fibre optic network for the conveyance of RF signals from a central controller to remote antennae at cellsite. This structure is particularly suitable to accomplish receiver and transmitter macrodiversity since the same RF channel can be transmitted and received through multiple ports, thus providing a large coverage area in which mobiles do not have to switch channels when roaming. This paper investigates the performance of multitransmitter diversity in fading channels, in noise and interference limited conditions. The problem differs from diversity reception in that signal combining is now non-coherent. Results are presented in terms of radio coverage and cochannel reuse efficiency firstly under Rayleigh fading only and secondly in a city street microcell environment. Simulcasting is an effective shadowing counteraction in noise conditions and achieves similar performance to no diversity in interference conditions.

We analyze the error floor of unequalized personal communications systems using minimum-shift keying (MSK) with training-sequence-based adaptive sampling in a two-delay fading channel. We include Gaussian filtering of the input, finite accuracy in determining the optimum sampling time, and filtering at the receiver. We prove that for pure MSK and low time dispersion, training-sequence-based adaptive sampling can completely avoid errors caused by intersymbol interference. The actual errors are caused by "secondary" effects (filtering and finite-resolution sampling) in conjunction with the channel time dispersion. Errors occur if the normalized phasor of the channel impulse response falls into certain error regions; the computation of these regions gives physical insights into the error mechanism and allows a highly efficient computation of the average bit-error rate (BER). The average BER varies as K/spl middot/(S/T)/sup 2/, where S is the RMS delay spread and T is the bit length. The proportionality constant K depends on the Gaussian filtering in the transmitter, the receiver filtering, and the amount of oversampling. The BER can be orders of magnitude lower than for the (quasi-) fixed sampling case, in which timing is derived solely based on the channel delay power profile. For two-branch diversity reception, we show that the BER is proportional to (S/T)/sup 4/.

In cellular mobile systems the on-line assessment of a radio link performance is normally associated with measurements of the received signal power, and is taken before signal demodulation. It determines the thresholds for handover policies, transmitting power control, diversity switching schemes, and other operations. However, a large received signal power does not necessarily mean "good quality" when it comes to considering radio links corrupted by cochannel interference, multipath propagation and background noise. In such cases, a better parameter that can be used to assess the signal quality is the signal-to-interference-plus-noise power ratio (SINR). Although this quality factor parameter is well-known in the literature, little has been written about how to measure it in a practical context. This article presents a method for the SINR estimation applied to a system using the Digital European Cordless Telecommunications (DECT) specification, but it can also be used by any other system employing constant modulus signals.

This paper presents an investigation into the effects of clipping of both baseband and bandpass multicarrier signals. It will be shown that considerable reductions in the amplitude variations (crest factor) of both classes of signal are possible at the expense of limited inband spectral distortion and some out of band clip noise emissions. The amount of inband distortion and out of band emissions can be traded off directly against reductions in the crest factor of the clipped signal. Novel clipping techniques are presented which give one more control over the various trade-offs.

The optimisation of resource management schemes is an important issue for future personal communications schemes. This paper begins with a discussion and analysis section on what the authors feel is the optimum performance for such a system employing resource management. A new approach to resource management is introduced called interference masking (IM), which differs from the usual approach, that of interference adaptation (IA). This method relies on accurate continuous power control with no lower bound on transmitter power so that interference becomes masked by the noise floor. The paper concludes with the reporting of results from comparing various combined power control and DCA schemes (some exploiting IM and some IA) using a comprehensive network simulation tool.

This paper reviews in summary form some of the contributions of COST-231 members to the study of the performance of the DECT system. Essentially, this has comprised three parts: radio link performance (and performance improvements via diversity and equalisation), propagation characterisation and network performance. The results reported here are concerned specifically with the limitations of performance due to the radio link in hostile environments, and the possible improvements via the use of diversity and/or equalisation.

High bit rate cordless transmission systems (e.g., DECT and emerging radio LANs) tend to use simple non-coherent receivers, and their performance can be constrained by channel dispersion. Equalisation can be used (supplementing other techniques such as diversity), but requires high frequency stability. In this contribution we use the DECT system as a case study and introduce a robust non-coherent receiver which performs equalisation at the output of a differential detector. The limitations of this basic structure are overcome by using a second, complementary differential processor. Simple algorithms are utilised to switch between the basic and complementary structures. The performance assessment of the new receiver is presented. The results suggest that the receiver achieves a significant extension of the dispersive channel operation range for DECT (maximum tolerable delay spreads of around 460-500 ns) with a multipath diversity gain comparable to a coherent equaliser receiver. On the other hand, unlike coherent structures, it remains largely unaffected by both frequency offsets and modulation index drifts.

Radio signals arriving from cellular mobile systems in urban areas can be modelled as a random process having the Nagakami-Rice probability density function. This assumption is made in order to design a receiver for combating the effects of multipath propagation for M-ary PSK modulated signals. A parameter based on the Nagakami fading index is explored and taken to construct a cost function to be minimized. The error surface is investigated and some strategies are examined for designing an equalizer. A modified constant modulus algorithm is presented and simulations are performed to test the efficiency of the proposed filter.>

The subject of the paper is an investigation to establish the minimum receiver complexity required when employing cells of different sizes. This was accomplished by using a GSM radio link simulation tool in conjunction with real channel data obtained by France Telecom's CNET in the centre of Paris, which was classified in terms of the distance between transmitter and receiver. The paper evaluates the trade-off between cell size and receiver complexity by simulating the performance of five alternative equalisers. In addition, the preprocessing operations necessary to import real channel sounding data into a radio link simulation are described in detail. It was found that when using a fixed height rooftop antenna, a simple 2 state receiver equaliser would only be acceptable if used in conjunction with additional multipath countermeasures. In terms of the current GSM system, the study shows that a 4 state equaliser can be used within about 1.2 km of the base station. Beyond this range, the standard GSM receiver is needed to provide the required performance.>

The multitone data transmission format has been receiving increasing interest recently as a potentially viable technique for transmission over both fixed and mobile links. However a disadvantage of the modulation format is that each transmitted multitone symbol has a non-constant envelope, with large instantaneous power spikes possible. This can lead to non-linear effects in the modulator causing intermodulation distortion and spectral spreading. In this paper the use of a peak power limit (or clip level) is considered for each symbol which can help to maintain linearity. The impact of applying such a clip level is analysed in terms of the symbol error rate (SER) caused on each subchannel, both theoretically and through simulation. Also a simple strategy for reducing the SER on each subchannel under a given clip constraint is introduced.>

A new structure for a non-coherent equaliser receiver applicable to the DECT-type systems is derived, where the equaliser algorithm takes into account the receiver non-linearities. It employs a modified channel estimation procedure suitable for differential detection. The outputs of the differential detector itself are then fed to a simple 2-state Viterbi algorithm which uses the modified channel estimates. Performance results are given for the new receiver which show that it can yield significant improvements with respect to standard non-coherent receivers in dispersive channels. At the same time, however, it will operate with receiver frequency offsets two orders of magnitude greater than a standard coherent detector equaliser structure (similar to that found in GSM) and is much less sensitive to the variations in the modulation index. In addition, performance limits of the basic structure are discussed and techniques for further improvement of the performance are considered.>

Previous work has shown that DCA offers a gain in system performance by adapting channel reuse to both the traffic and interfering environment. In this paper a third form of adaptation (location adaptation) is proposed and analysis is used to predict the potential gain this would provide. A simulation tool is used to investigate the performance of some DCA schemes with reference to the calculated potential gain. Two simple distributed algorithms are compared with a computationally intensive global DCA method. The simulation results suggest that if channel reassignment is used when call quality drops below an acceptable level, certain distributed DCA schemes may outperform global techniques. The results show how the reassignment process organises the system towards an optimised situation but that some form of additional interference adaptation is required to achieve the optimum.>

It is pointed out that in the paper, the authors did not take into account the effect of deemphasis on the received signal power when calculating the SNR for sinusoidally modulated SSB-FM.