This paper elucidates through experiments the improvement in the achievable bit error rate (BER) performance when space time transmit diversity (STTD) is applied to the wideband direct sequence code division multiple access (W-CDMA) forward link. First, laboratory experimental results clarify that the received path timing difference of transmitted signals from two antennas, due to the propagation delay, should be within a chip duration of approximately 1/4 and 1/2 with and without fast transmit power control (TPC), respectively, in order to achieve a prominent transmit diversity effect. We show that the required average received signal energy per bit-to-background noise spectrum density (Eb/N0) at the average BER of 10-3 using STTD is decreased by approximately 4.2 (1.7) dB compared to the case of single-antenna transmission at the maximum Doppler frequency, fD, of 5 Hz without (with) antenna diversity reception at a mobile station (MS) due to the increasing randomization effect of burst error. Furthermore, we elucidate that although the gain of STTD in field experiments is decreased compared to that in laboratory experiments, since the degradation in path search accuracy is greater due to the frequently changing delay time of each path in a real multipath-fading channel, the required average received signal energy per bit-to-interference plus background noise power spectrum density ratio (Eb/I0) at the average BER of 10-3 with STTD is decreased by approximately 1.3 to 1.5 (0.7 to 1.0) dB without (with) antenna diversity reception when fast TPC is not applied in the forward link. This indicates that STTD is effective for a channel without TPC such as a common control channel in a real multipath-fading channel.

This paper proposes an inter-cell site diversity scheme based on 2-step selection combining (SC) and investigates through experimentation the effect of inter-cell site diversity in the transmit power-controlled wideband direct sequence code division multiple access (W-CDMA) reverse link. In the proposed algorithm, the decoded data sequence after soft-decision Viterbi decoding at each base station (BS) is transferred via the backhaul (wired line between BS and radio network controller (RNC) simulator) to the RNC simulator accompanied by reliability information of cyclic redundancy check (CRC) results per frame and the average signal-to-interference power ratio (SIR) calculated over the interleaving interval. The 2-step SC for each frame is performed at the RNC simulator using these two types of reliability information. We conclude from the laboratory experiments that the transmit power of a mobile station (MS) can be decreased since the selection period, TSEL, is shorter irrespective of the interleaving length, TILV, and that the required transmit power of a MS satisfying the average BER of 10-3 in inter-cell site diversity among 2 and 3 cell sites can be decreased by approximately 1.0 (0.5) dB and 1.3 (0.7) dB for fading maximum Doppler frequency fD = 5 (80) Hz, respectively, compared to a one-site connection (TILV = 80 msec, TSEL = 10 msec, path loss difference between 2 BSs and MS is 0 dB). We also confirmed by field experiments that the required transmit power of a MS in inter-cell site diversity between 2 cell sites can be decreased by approximately 2.0 dB compared to that of a one-site connection.

This paper investigates the optimum adaptive antenna array beam forming (AAA-BF) configuration considering the diversity effect provided by transmit diversity (TD) in a multipath fading channel in the W-CDMA forward link. Computer simulation results show that the required transmit signal energy per bit-to-background noise power spectrum density ratio (Eb/N0) at the average block error rate (BLER) of 10-2 using AAA-BF coupled with TD is decreased by approximately 1.0 dB compared to that of AAA-BF assuming the identical number of total antennas when the capacity, i.e., the number of simultaneously accommodated users with the transmission power proportional to the symbol rate, normalized by processing gain, Pg, is below approximately 20%. However, we find that in an interference-limited channel, when the capacity normalized by Pg is above approximately 30%, AAA-BF employing all antennas accommodates a larger capacity compared to AAA-BF coupled with TD because of a sufficient interference suppression effect due to a much narrower beam width despite the absence of the antenna diversity effect. This paper also elucidates in a multi-cell model that AAA-BF employing all antennas can accommodate approximately 1.5 times more users with the average BLER of 10-2 and with the outage probability of 5%, compared to the case with AAA-BF coupled with TD, when the total number of antennas is 8.

A RAKE combiner based on a matched filter (MF) can be relatively easily implemented since the despread signal components that have propagated along different paths appear sequentially at the MF output. An important design problem is how to accurately select the paths having sufficiently large signal-to-noise power ratios (SNRs). This paper proposes a simple path selection algorithm that uses two selection thresholds. The first threshold is to select the paths that provide largest SNRs. However, as the total received signal power (sum of the signal powers of all paths) decreases, some of the selected paths become noisy. Therefore, we introduce a second threshold that discards the noisy or noise-only paths from among those selected by the first threshold. We apply the proposed path selection algorithm to a pilot symbol-assisted coherent RAKE combiner and find by computer simulations a near optimum set of the two thresholds in frequency selective multipath Rayleigh fading channels. Several power delay profile shapes are considered. The simulation results demonstrate that the MF-based RAKE combiner with the two selection thresholds can achieve a bit-error-rate (BER) performance close to the ideal case (i. e. , the paths to be used for RAKE combining are selected for each power delay profile such that the required signal energy per information bit-to-noise spectrum density ratio (Eb/N0) is minimized).

In estimating the performances of Distributed control Dynamic Channel Assignment (DDCA) strategies in sector cell layout systems, we find that sector cell layout systems with DDCA achieved a large system capacity. Moreover, we also indicate the problem, which is the increase of occurrences of cochannel interference, raised by using DDCA in sector cell layout systems. The new channel assignment algorithm, which is called Channel Searching on Direction of Sector (CSDS), is proposed to cope with the problem. CSDS assigns nominal channels to each sector according to their direction so that the same frequency channel tends to be used in sectors having the same direction. We show, by simulations, that CSDS is an adequate algorithm for sector cell layout systems because it significantly improves performance on co-channel interference while only slightly decreasing system capacity.

This paper evaluates the effect of inter-sector diversity with maximal ratio combining (MRC) coupled with coherent Rake combining and 2-branch antenna diversity reception in the transmit-power-controlled wideband direct sequence code division multiple access (W-CDMA) reverse link. We first elucidate based on laboratory experiments that the required average transmit signal energy per bit-to-background noise spectrum density ratio (Eb/N0) at the average bit error rate (BER) of 10-3 with inter-sector diversity using two sectors is decreased by approximately 1.4, 1.0, and 0.2 dB compared to that with inter-cell site diversity using two cell sites with antenna diversity reception due to the superiority of MRC to selection combining (SC), when the difference in the average path loss between a base station (BS) and a mobile station (MS) is Δ12 = 0, 3, and 6 dB, respectively. We also clarify in actual field experiments that the inter-sector diversity associated with Rake time diversity and antenna diversity further decreases the required average transmit power of a MS if the number of resolved paths is small such as 1 or 2 in each sector reception, even when the fading correlation between sectors is relatively large. Furthermore, we show that the required average transmit power of a MS for satisfying the average BER of 10-3 with inter-sector diversity is decreased above approximately 2.0-2.5 dB compared to that with one-sector reception, owing to the significantly increased inter-sector diversity effect in addition to the Rake time diversity and antenna diversity, when the fading correlation averaged over the measurement course is approximately 0.7.

This paper experimentally demonstrates the possibility of 2-Mbps data transmission using a 5-MHz bandwidth (chip rate of 4.096 Mcps) wideband DS-CDMA (W-CDMA) mobile radio link in frequency-selective multipath fading environments. To reduce the mobile station transceiver complexity, three-orthogonal code multiplexing with the spreading factor (SF) of 4 is employed. In such a small SF transmission, the increased multipath interference (MPI) significantly degrades the transmission performance. We consider two-branch antenna diversity reception and fast transmit power control (TPC) as well as channel coding to mitigate the influence of MPI. Laboratory experimental results show that the use of antenna diversity reception is significant and that the fast TPC improves the transmission performance. Furthermore, the impact of the fading maximum Doppler frequency, fD, and that of the channel coding interleaving size, Tint, on the achievable BER performance are also investigated.

In DS-CDMA (including W-CDMA), a received signal can be resolved into multiple paths to be Rake combined. An important design problem of the Rake receiver is how to accurately search the paths with a sufficiently large signal-to-interference plus background noise power ratio (SIR). This paper investigates the performance of a coherent Rake receiver using pilot symbol-assisted channel estimation with fast transmit power control, and thereby optimizes three key parameters: the total averaging period, Tavg, consisting of a combination of coherent summation and power summation; each period of the summations for measuring the average power delay profile; and path-selection threshold M from the generated power delay profile. We used a path search algorithm, which searches the paths that have M times greater average signal power than the interference plus background noise power measured in the average power delay profile generated using time-multiplexed pilot symbols. It was clarified by both simulation and laboratory experiments that when M = 4, Tavg = 50-100 msec, and the number of slots for coherent accumulation R = 2, the required average transmit Eb/N0 for obtaining the average BER of 10-3 is almost minimized with and without antenna diversity for both ITU-R Vehicular-B and average equal power L-path delay profile model, in which each path suffered independent Rayleigh fading. The paper also shows that based on the field experiments, the path search algorithm with optimized path-selection parameters is robust against actual dynamic changes in the power delay profile shape.

This paper evaluates the bit error rate (BER) performance of high rate data transmission such as at 64 and 384 kbits/s (kbps) with high quality (average BER is below 10-6) using turbo/convolutional coding associated with Rake time diversity, antenna diversity, and fast transmission power control (TPC) in multipath fading channels for W-CDMA mobile communications. Laboratory experiments using multipath fading simulators elucidate the superiority of turbo coding over convolutional coding when the channel interleaving length is 40 msec. The required average transmission power for the average BER of 10-6 using turbo coding is decreased by approximately 1.1-1.5 dB and 1.5-1.6 dB for 64 and 384 kbps data transmissions, respectively, compared to that using convolutional coding for a two-path Rayleigh fading channel with the fading maximum Doppler frequency of fD = 5-200 Hz. Furthermore, field experimental results elucidate that the required transmission power for the average BER of 10-6 employing turbo coding is decreased by approximately 0.6 dB and 2.0 dB compared to convolutional coding for 64 and 384 kbps data transmissions, respectively, without antenna diversity reception, while that with antenna diversity reception exhibits only an approximate 0.3-0.5 dB decrease. This decrease in improvement with antenna diversity reception indicates that in an actual fading channel in the field experiments, the impact of the error in path search for Rake combining and SIR measurement for fast TPC diminishes the performance improvement of the turbo coding due to a very low received signal power.