This paper illustrates the performance of the high chip rate UTRA-TDD system deployed in an indoor office type environment. User terminals (or UEs) are distributed randomly within an office building that is served by 10 TDD base stations (BSs). Most terminals will generate 12.2kbit/s traffic but a small proportion of these generate 64kbit/s and 144kbit/s traffic. The results obtained by Monte Carlo simulation methods show (a) range of the served terminals, (b) path loss distributions for the served terminals, (c) inter-cell and intra-cell interference distributions, (d) capacity supported per base station and (e) coverage plots demonstrating the probability of outage at each office location. The TDD UL supports more capacity than the DL per radio resource.

In the universal mobile telephony system (UMTS) the frequency division duplex (FDD) and time division duplex (TDD) modes have adjacent carriers at 1920 MHz. This creates adjacent channel interference (ACI) between the two different air interfaces. Since different duplexing modes are used, the implications for each system are different, with respect to capacity and coverage: these implications are investigated in the paper. The separation distance of the TDD and FDD base station and the load in each system are varied and a symmetrical speech service in both systems is considered, with non-ideal power control assumed. It is found that for an FDD cell radius of 1000 m, a TDD cell radius of 50 m and 10% maximal tolerable outage, the effects of ACI on capacity can be compensated by dynamically increasing the required power at each BS, without affecting the coverage.

We present two types of multiple input channel models that have been implemented: a geometric model and a stochastic model. We consider the relative merits and limitations of these models. There are a number of factors that will indicate whether geometric or stochastic multichannel models are appropriate for a particular test scenario. The advantage of a geometric model is that all of the factors used to characterise the channel (e.g. Doppler spread, delay spread, angle of arrival, etc.) are correlated. The geometric model, by its physical ray tracing nature, automatically implements these artefacts of the signal and can be used in indoor picocells and outdoor macrocells without the need for empirical data. The disadvantage of the geometric model is that it is a simplification of the physical world, and so some of the statistical detail is likely to be lost

The Universal Mobile Telecommunications System (UMTS) is composed of an FDD and TDD mode. Both modes have an adjacent carrier at 1920 MHz. The implications thereof for system capacity are analysed. In this context, the FDD mode Is assumed to account for full spatial coverage and the TDD mode is employed to cover areas with increased traffic 'hot spots'. The separation distance of the TDD and FDD base station (BS) and the load in each system are varied. It is shown that the capacity in the FDD system is reduced by less than 5 % for a BS separation distance greater than 300 m. Contrary to this, the capacity reduction in the TDD system is only below 5 % for a BS separation of less than 1000 m, This clearly yields a tradeoff which limits the degree of freedom in planning the network in this particular case.

When comparing UTRA-TDD with UTRA-FDD it can be found that in the UTRA-TDD mode additional interference scenarios exist. Mobile stations (MS's) can interfere with each other and so can base stations (BS's), Since the source and sink of this type of interference are the same we call it same-entity interference. It is shown by a novel interference resolving algorithm that same-entity interference can be constructively exploited to enable asynchronous overlaps in a UTRA-TDD network, Asynchronous overlaps exist when any cell A is transmitting while the neighbour cell B is receiving. It was found that in a network where an asynchronous overlap exists the algorithm proposed reduces outage from 14 % to 6 %. In contrast, the outage of an ideally synchronised network was found to be 3.5 %. In this 'ideal' network asynchronous overlaps are disallowed resulting in a significant drawback. This is that the flexibility of a TDD system to easily adopt different channel asymmetries is significantly limited.

The Universal Mobile Telecommunications System (UMTS) is composed of an FDD and TDD mode. The spectrum allocation is in such a way that both modes have an adjacent carrier at 1920 MHz. The implications thereof with respect to system capacity are investigated. In this context the separation distance of the TDD and FDD base station (BS) and frame synchronisation are varied. It is shown that the most detrimental effects for the FDD interface are for small BS separations. In contrast, the optimum with respect to TDD capacity is found for the co-location of both base stations. This yields a trade-off for the optimal BS locations. Moreover, an adjacent channel protection factor of 30 dB in an interference limited system such as CDMA is shown to be too low unless a significant capacity loss is acceptable.

A new method for initial code acquisition in spread spectrum systems is compared in terms of the resulting mean acquisition time with the single dwell acquisition scheme in the case where a non uniform a priori information is provided and advanced searching strategies are utilised. The novel scheme employs a fast preliminary search of the ambiguity area. The results of the fast search are used as a priori information for the final search. The probability of false ranking P-fr is defined as a measure of thr reliability of the a priori information provided by the fast preliminary search. A similar idea, used to perform a spatial search for a lost satellite in a region of the sky, has been described in [1]. A detailed analysis of the FPS system is presented in [2].

The UMTS terrestrial radio access (UTRA) is composed of a frequency division duplex (UTRA-FDD) mode and a time division duplex (UTRA-TDD) mode. In UTRA-TDD the uplink and downlink are on the same carrier frequency This creates additional interference scenarios, to be precise: MS <----> MS and BS <----> BS interference. Since power leakage between adjacent carriers cannot be avoided this inherent property of TDD has an impact on adjacent channel interference (ACI) and thus on cell capacity. The power leakage results from transmitter mask imperfections and non-ideal receiver filters. From interference power measurements using a 7th and a 9th order butterworth receiver filter the relationship between carrier spacing and adjacent channel interference power ratio (ACIR) is established. This relationship is used to find the best carrier spacing in UTRA-TDD considering several cell deployment scenarios. It is found that the ACIR must be greater than 30 dB, but the capacity gains from an ACIR greater than 40 dB converge rapidly.

We investigated the feasibility of the coexistence of CDMA-TDD micro cells under a CDMA-FDD macro cell network with both sharing the same frequency band. A dynamic channel allocation (DCA) algorithm decides on the basis of mutual interference which FDD band is to use for the micro cell TDD channel. On the assumption of giving the macro cell higher priority we investigated the lower bound probability of total micro cell loss (outage) dependent on the distance between the micro- and macro cell BS. We also considered wall loss at the boundary of the micro cell and studied the effect of different propagation conditions on the assumption of equally distributed macro cell subscribers. A mathematical model for outage is developed and the results are compared with simulations. The results reveal a good match between the theoretical approximation and the simulations. This, in turn, indicates that micro cell outage is mainly caused by a single macro cell mobile within a certain area around the micro cell BS. It can be seen that the outage goes up to about 15% when the micro cell is located at the macro cell boundary. This value exponentially decreases when the micro cell is moved towards the macro cell BS. Walls around micro cells do have a significant influence on outage. It has been shown, that outage converges towards a value determined by the micro cell size.

A new receiver architecture for PN code acquisition of signals with high Doppler shifts is proposed, in which the partially correlated outputs of the parallel section of a serial-parallel correlator (SPC) are subject to an FFT process before being summed by the serial section. In this way a wider range of frequencies can be searched in a single dwell time, which is particularly useful in a LEOS channel due to the high Doppler shifts involved. The performance in terms of mean acquisition time is shown to be better over a wider bandwidth than either serial or parallel correlators.

The authors describe a robust channel prediction technique for a direct sequence spread spectrum (DS-SS) system in a fast fading environment. For improved performance the RAKE filter taps are coherently combined, hence accurate channel estimation is required. An FIR type linear prediction filter for each RAKE filter tap is used to estimate the channel response. In order to do this, the data decisions are fed back to the prediction filter. The stability of the proposed system is achieved through differential encoding of the data bits. It is demonstrated through simulations that the performance of the proposed decision directed receiver is better than that of an idealised receiver where channel estimation is not corrupted by decision feedback errors (e.g. by means of employing a pilot signal). The channel estimate can be significantly improved by employing a second stage channel estimation filter.

The basic concepts of a time division duplex system are introduced. The advantages of using TDD-CDMA for mobile radio communications is shown to be related to its ability to handle asymmetric data and benefits resulting from the reciprocal nature of the channel. Synchronisation difficulties and the associated interference problems are seen as the primary limiting factors. The use of TDD-CDMA in UMTS is highlighted and the proposal for sharing FDD bandwidth allocations with the TDD system is explained and outage problems are identified.

The problem of coherent detection for a diversity receiver to combat the severe penalty of transmitting over a Rayleigh fast fading channel is addressed. The principle of per-survivor processing is employed to estimate the channel response for BPSK modulated signals. To achieve this, a symbol aided plus decision directed transmission scheme is used. Simulation results demonstrate that the proposed algorithm can track the fading fluctuations

The authors describe a robust channel prediction technique for a direct sequence spread spectrum (DS-SS) system in a fast fading environment. Coherent diversity reception is employed for improved performance. To estimate the channel response, an FIR type linear prediction filter is employed for each RAKE tap. The stability of the decision directed receiver is achieved through differential encoding of the data bits. It is demonstrated through simulations that the performance of the proposed decision directed receiver is better than a receiver which relies on a pilot signal.

The performance of an OFDM-CDMA system with convolutional coding and interference cancellation is presented. The OFDM-CDMA receiver uses equal gain combining on the carriers before the received signal is passed to the interference canceller. Within the interference canceller, the interference estimates are improved by Viterbi decoding before being subtracted from the wanted signal. Both orthogonal and super orthogonal coding are considered and their performance is compared against a 3/4 rate K=7 punctured convolutional code. A comparison is made of the spectral efficiencies of the different schemes.

The performance of different channel coding schemes are presented for an OFDM-CDMA system. The different channel coding schemes are shown for a receiver using equal gain combining on the carriers in a multipath channel. As such the only information needed at the receiver is the phase of the sub-carriers. Two new channel coding schemes for OFDM-CDMA are presented, orthogonal and super orthogonal convolutional coding. These are compared against a 1/2 rate K=7 convolutional code and a 3/4 rate K=7 punctured convolutional code. A comparison is made of the spectral efficiency of the different coding schemes, the bandwidth required and the receiver complexity tradeoffs.

This paper considers a time division duplex (TDD) extension applied to a conventional frequency division duplex (FDD) CDMA cellular system. The FDD-CDMA cellular system is overlaid by small TDD-CDMA cells which can cover capacity "hot-spots" such as in offices and public buildings. The TDD-CDMA base stations have lower complexity than the FDD-CDMA base stations and are able to support asymmetric traffic at high bit rates.

The advantages of using a time division duplex-code division multiple access (TDD-CDMA) air interface for mobile multimedia services are presented, with reference to an equivalent frequency division duplex (FDD) CDMA system. In particular the flexibility and limitations of using an asymmetric channel are discussed. It is concluded that TDD-CDMA is ideally suited to short range high bit rate asymmetric services, whereas FDD-CDMA is better suited to large cells using symmetric channels.

A non-coherent fast acquisition system is described for a DS-SS signal in the presence of a large Doppler shift. This technique utilises N complex correlators of length x such that Nx=m where m is the length of the spreading sequence. These complex correlators produce partial correlations of the DS-SS signal. The outputs of these correlators are fed to an N point FFT. The maximum output of the FFT is then selected and compared against a threshold. In situations with high Doppler shifts the FFT acquisition system described can test all frequency offsets for a given code phase simultaneously. The theoretical probability of detection and false alarm are derived in a Gaussian channel. We determine the mean acquisition time for an 8-point FFT using a single dwell system in the presence of 32 kHz Doppler shift. It is shown by choosing a suitable threshold level that the mean acquisition time is over 16 times lower than a standard single dwell non-coherent decorrelator.

The use of RAKE receivers in systems which are subject to rapidly fading multipath propagation requires an adaptive algorithm with a predictive element to avoid large errors due to phase lag. A novel use of Prony's (1795) method, as a predictive algorithm in RAKE receivers, is proposed, and its performance in terms of bit error rate evaluated using the COST 207 model of a typical urban environment. It has been found that a predictive RAKE receiver using the Prony algorithm performs particularly well at high Doppler frequencies as compared with a similar receiver using alpha-trackers to estimate the tap amplitudes and phases