This paper proposes a bit-selective iterative processing scheme for turbo equalization. We consider convolutionally coded and turbo coded transmissions over inter-symbol interference (ISI) channels. The receiver starts with conventional iterative channel equalization and decoding. After each iteration, bit convergence status is measured via cross-entropy. Short windows are then applied around unconverged bits and the subsequent detection/decoding iterates only within the windows. The rapid decrease in selected bits after each such iteration reduces computational complexity significantly, but performance is well maintained by utilizing the Markov property in the trellis detection/decoding.

Near-capacity performance of turbo codes is generally achieved with a large number of decoding iterations. Various iteration stopping rules introduced in the literature often induce performance loss. This paper proposes a novel partial decoding iteration scheme using a bit-level convergence test. We first establish decoding optimality of windowed partial iteration for non-converged bits given that convergence has been achieved on window boundaries. We next present two criteria for testing bit convergence based on cross-entropy, and propose a windowed partial iteration algorithm. The overall complexity and memory requirements of the new algorithm are evaluated and compared with known algorithms. Simulations reveal that the proposed scheme suffers essentially no performance loss compared to full iterations, while reducing the decoding complexity. We also briefly discuss possible extensions of the proposed scheme to general iterative receivers.

Multicarrier techniques based on affine Fourier transform (AFT) have been recently proposed for transmission in the wireless channels. The AFT represents a generalization of the Fourier and fractional Fourier transform. We derive the exact and approximated interference power, upper bound and measure of applicability for the AFT based multicarrier (AFT-MC) system. It is demonstrated that the AFT-MC system effectively minimizes interference in time-varying multipath channels with line-of-sight component and narrow beamwidth of scattered components that often occurs in aeronautical and satellite communications.

Based on an analysis on the recursive computation of the iterative maximum a posteriori (MAP) algorithm for turbo decoding, this paper considers a modified MAP scheme with reduced block lengths for symbols with unreliable detection after some initial iterations. Applying symbol selection based on cross-entropy measurement for parallel concatenated convolutional codes, we develop partial, windowed iterations for selected symbols. By omitting computations for symbols with reliable detection results, this approach significantly reduces complexity but well maintains the performance by complete iterations.

Cooperative transmission, in which a source and relay cooperate to send a message to a destination, can provide spatial diversity against fading in wireless networks. We derive analytical expressions for the error probability of amplify-and-forward (AF), decode-and-forward (DF), and a new hybrid AF/DF relaying protocol, for systems using strong forward error correction in quasi-static Rayleigh fading channels, and these expressions are shown to compare favorably with simulation results using turbo codes. Analytical results include an exact expression for the distribution of the SNR in AF transmission. For the protocols that achieve diversity (AF, adaptive DF, and hybrid AF/DF), the optimum position of the relay is midway between the source and destination, implying that mutual relaying (or partnering) to a common destination is suboptimal

This paper investigates the benefit of adaptive modulation based on channel state information (CSI) in direct-sequence/ code-division multiple-access (DS/CDMA) multihop packet radio networks. By exploiting varying channel conditions, adaptive modulation can be used in ad hoc networks to provide upper layers with higher capacity links over which to relay traffic. Using the /spl alpha/-stable interference model, the distribution of the signal-to-interference ratio (SIR) is obtained for a slotted system of randomly, uniformly distributed nodes using multilevel coherent modulation schemes. Performance is evaluated in terms of the information efficiency, which is a new progress-related measure for multihop networks. Three types of adaptivity are analyzed, differing in the level of CSI available: 1) full knowledge of the SIR at the receiver; 2) knowledge of only the signal attenuation due to fading; and 3) knowledge of only the slow fading component of the signal attenuation. The effect of imperfect channel information is also investigated. Sample results are given for interference-limited networks experiencing fourth-power path loss with distance, Ricean fading, and lognormal shadowing.

We study a cellular network with one multiantenna base station and a number of multiantenna users. Correlated fading may exist at the base station or at the users, or at both sides. With time-varying fading, multiuser diversity is exploited by always allowing the base station to transmit to the user with the best channel. For this network, we find that if the number of transmitter antennas and the number of receiver antennas grow with a fixed ratio, the multiuser diversity gain will approximately remain constant

In this paper, we study the uplink of a cellular network where reservation random access schemes such as reservation slotted ALOHA and reservation slotted nonpersistent ISMA are used for multiple access. When channel state information is not available to users when they contend for channel access, we find that capture effect inherent in wireless communications can exploit the channel variations and reserve users with favorable channels for data transmission. We investigate two channel models where in the first model, channel variations come from time-varying Nakagami-m fading and in the second model, mobile users have both time-varying fading and path loss

Previous studies have shown that single-user systems with t antennas at the transmitter and r antennas at the receiver can achieve capacity proportional to min(r, t) assuming a rich scattering environment. In this paper, we study a cellular network with a t-antenna basestation and K r-antenna users. We assume exponentially correlated fading exists either at the basestation or at the users, and the channels of different users change independently. Multiuser diversity is exploited by always allowing the basestation to transmit to the user which channel provides the highest capacity among all users. For this network, we investigate the impact of correlated fading on the network capacity. We find that if both t and r > 1, the impact of correlated fading depends on the signal-to-noise ratio (SNR). At the low SNR regime, the capacity will also increase with the correlation coefficient. However, at the high SNR regime, under the condition that one side of the channel having correlated fading has less antennas than the other side, the increase of the correlation coefficient will reduce the capacity. In addition, we find the closed-form solutions for the mean and variance of single-user multiple-input multiple-output (MIMO) channel capacities for different cases.

In this paper, we evaluate the correlation coefficient between two multiple-input multiple-output (MIMO) channel capacities when these channels have temporal correlations. We find the exact value of the correlation coefficient for low signal-to-noise ratio (SNR), and its upper/lower bounds and approximation for high SNR. The obtained results have important applications such as calculating the level crossing rate

This paper derives new soft decision metrics for coded orthogonal signaling in symmetric /spl alpha/-stable noise, which has been used to model impulsive noise. In addition to the optimum metrics for Gaussian (/spl alpha/=2) noise and Cauchy (/spl alpha/=1) noise, a class of generalized likelihood ratio (GLR) metrics with lower side information requirements is derived. Through numerical results for a turbo code example, the Cauchy decoder is found to be robust for a wide range of /spl alpha/, and GLR metrics are found which provide performance gains relative to the Gaussian metric, but with lower complexity and less a priori information.

The paper presents a novel cooperative coding scheme using punctured serial concatenated convolutional codes (SCCC). We evaluate its performance in terms of frame error rate (FER) under two cooperation protocols: a decode-and-forward protocol in conjunction with automatic repeat request (ARQ), by which the relay sends an acknowledge bit (ACK) back to the source to indicate whether its data packet is correctly decoded; an adaptive amplify-and-forward protocol, where the amplifying relay node cooperates to forward the message only when the inter-user link gain is above a certain threshold. In both cases, the destination makes the decision on each transmitted bit after appropriately combining signals and performing iterative decoding. Simulation results show that SCCC-based cooperative coding achieves full spatial diversity and exhibits a similar FER performance to that using parallel concatenated convolutional codes (PCCC). At the price of channel feedback, the adaptive decode-and-forward algorithm demonstrates a relatively lower FER than the adaptive amplify-and-forward scheme. Amplified cooperation experiences severe performance degradation due to the inefficiency of repetition coding and error propagation, especially when the quality of the source-relay link deteriorates; however, if the threshold for carrying cooperation is carefully selected, the adaptive amplify-and-forward protocol can perform comparably to the decode-and-forward protocol.

This paper analyzes the benefit of adaptive routing based on knowledge of the channel state information in multihop, ad hoc wireless networks that use direct-sequence code-division multiple access. Cross-layer, channel-adaptive routing exploits the inherent spatial diversity of multihop wireless networks to select links with favorable channel conditions. The information efficiency, an extension of a previously used measure called expected progress, is used to evaluate performance. Results show that, combined with adaptive modulation, adaptive routing can improve performance in ad hoc networks by a factor of four to five in channels with Rayleigh fading and lognormal shadowing. The lack of position information in the routing decision would reduce performance by 25%. New approaches to channel-adaptive routing that enable rapid adaptivity to channel conditions are discussed.

In a cooperative transmission system, a source cooperates with a relay to transmit a message to a local destination. This paper proposes an adaptive cooperative protocol utilizing turbo codes and studies the trade-off between spatial and temporal diversity in increasingly mobile channels. While the advantage of cooperation is most pronounced in slow fading channels—on the order of 8-12 dB for the system considered here—the benefit gradually decreases with increasing mobility and vanishes at a normalized Doppler spread of 10–2. Nevertheless, for typical wireless packet data services, cooperative coding can provide appreciable gain even in moderately mobile environments (e.g., up to 50 km/h). While the protocol implicitly adapts to varying channel conditions, adaptivity based on explicit knowledge of channel state information is also considered, and we find that the adaptive choice of two relays reduces the power requirement in half.

Cooperative diversity transmission exploits the spatial diversity created by antenna sharing in a wireless network. The objective of this paper is to investigate the cooperative diversity technique performed in a multi-hop DS-CDMA system, where the limiting factor in the system is the multi-access interference (MAl). We focus on the use of minimum mean-square error (MMSE) multiuser detection (MUD) in the uplink of cooperative CDMA networks. MMSE detectors are designed to suitably combine the signals from the direct path and relaying path and minimize the signal in the interference-and-noise ratio at the receiver output. We evaluate the bit error rate (BER) performance of cooperative transmission in narrow-band block fading channel. Also, MMSE detection is applied in a random CDMA network and the average performance is presented through Monte-Carlo simulations. It is illustrated that, through using MMSE detection in a non-orthogonal CDMA system, cooperation is beneficial in terms of introducing additional diversity gain and thus ensuring a low bit error rate.

In this paper, as part of the adaptive resource allocation and management (ARAM) system (Alagoz, 2001), we propose an adaptive admission control strategy, which is aimed at combating link congestion and compromised channel conditions inherent in multimedia satellite networks. We present the performance comparisons of a traditional (fixed) admission control strategy versus the new adaptive admission control strategy for a direct broadcast satellite (DBS) network with return channel system (DBS-RCS). Performance comparisons are done using the ARAM simulator. The traffic mix in the simulator includes both available bit rate (ABR) traffic and variable bit rate (VBR) traffic. The dynamic channel conditions in the simulator reflect time variant error rates due to external effects such as rain. In order to maximize the resource utilization, both for fixed and adaptive approaches, assignment of the VBR services are determined based on the estimated statistical multiplexing and other system attributes, namely, video source, data transmission, and channel coding rates. In this paper, we focus on the admission control algorithms and assess their impact on quality-of-service (QoS) and forward link utilization of DBS-RCS. We show that the proposed adaptive admission control strategy is profoundly superior to the traditional admission control strategy with only a marginal decrease in QoS. Since the ARAM system has several parameters and strategies that play key roles in terms of the performance measures, their sensitivity analysis are also studied to verify the above foundations.

This paper investigates the benefit of adaptive modulation based on channel state information (CSI) in DS/CDMA multihop packet radio networks. By exploiting varying channel conditions on different links, adaptive modulation can be used in ad hoc networks to provide upper layers with higher capacity links over which to route traffic. Performance is evaluated in terms of the information efficiency, a new progress-related measure for multihop networks. Three types of adaptivity are analyzed, differing in the level of CSI available: (i) full knowledge of the SIR at the receiver, (ii) knowledge of only the signal attenuation due to fading, and (iii) knowledge of only the slow fading component of the signal attenuation. The effect on performance of imperfect channel information is also investigated. Sample results are given for interference-limited networks experiencing fourth-power path loss with distance, Rayleigh fading, and lognormal shadowing.

We propose a new suboptimum multiuser detector for synchronous and asynchronous multiuser communications. In this approach, a greedy strategy is used to maximize the cost function, the maximum-likelihood (ML) metric. The coefficients of the ML metric are utilized as weights indicating in which order bits can be estimated. The complexity of the algorithm is approximately K/sup 2/ log K per bit, where K is the number of users. We analyze the performance of the greedy multiuser detection in the additive white Gaussian noise channel as well as in the frequency-nonselective Rayleigh fading channel, and compare it with the optimum detector and several suboptimum schemes such as conventional, successive interference cancellation, decorrelator, sequential, and multistage detectors. The proposed greedy approach considerably outperforms these suboptimum schemes, especially for moderate and high loads in low and moderate signal-to-noise ratio regions. The results show that when there is a significant imbalance in the values of the coefficients of the ML metric due to moderate to high noise, fading, and asynchronous transmission, near-optimum performance is achieved by the greedy detection.

Recently, a novel scheme for iterative multiuser detection and turbo decoding was proposed by the authors (2000). It has been shown that the proposed detector approaches the multiuser capacity limit within 1 dB in the low signal-to-noise ratio region. However, the main drawback of the scheme is its exponential complexity in the number of users, which is due to the complexity of the MAP multiuser detector. In this paper, we show that the complexity of the algorithm can be significantly reduced by employing a greedy multiuser detector. The results for both the additive white Gaussian noise and the single-path fading channels show that the substantial reduction in complexity introduces a slight degradation in performance, 0.5 dB compared to MAP detector.

Previous results have demonstrated excellent performance of turbo coded modulation with iterative demodulation-decoding for systems with antenna diversity. However, the computational complexity at the receiver is exponential in the number of transmit antennas; hence it is too high for the case of large number of transmit and receive antennas. We present a new iterative-greedy demodulation-decoding algorithm for turbo codes for multiple antenna systems. The decoding algorithm is based on a novel greedy detection method for multiuser communications. The iterative-greedy demodulation-decoding algorithm is characterized by the exchange of soft information between the demodulator and the decoder, and the significant decrease in computational complexity, as compared to the standard method of iterative demodulation-decoding, with a slight degradation in performance. This is due to the novel, greedy method of calculating the log-likelihoods to be used in the iterative decoding process, where only a small subset of the possible received vectors is considered, resulting in a receiver with only polynomial complexity in the number of transmit antennas.

Previously, a novel scheme for iterative multiuser detection and turbo decoding was proposed by Damnjanovic and Vojcic (2000, 2001). In this scheme, multiuser detection and single-user turbo decoding are tightly coupled to maximize the overall gain. The extrinsic probabilities for the coded bits of the interfering users, obtained after each turbo decoding iteration, are used as a priori probabilities in the following multiuser iteration and the extrinsic information for the systematic bits of the desired user is used as a priori information in the next single-user turbo decoding iteration. Turbo decoding of parallel concatenated convolutional codes is carried out in parallel fashion. It has been shown that the proposed detector approaches the multiuser capacity limit within 1 dB in the low signal-to-noise ratio region. However, the main drawback of the scheme is its exponential complexity in the number of users, which is due to the complexity of the maximum a posteriori probability (MAP) multiuser detector. In this paper, we show that the complexity of the scheme can be significantly reduced by replacing the (MAP) multiuser detector with an iterative detector derived from the greedy multiuser detector proposed by AlRustamani and Vojcic (2000). In this paper, we demonstrate that, for both the additive white Gaussian noise and the frequency-nonselective Rayleigh fading, the substantial reduction in complexity of the iterative scheme proposed by Damnjanovic and Vojcic when the greedy detector is utilized introduces a slight degradation in performance.

In this paper, we consider coded asynchronous multiuser signals in an additive white Gaussian noise channel. Since optimum joint multiuser detection (MUD) and forward error correction (FEC) decoding is characterized with a very high computational complexity, we consider disjoint MUD and FEC decoding. The optimum disjoint multiuser detector is the soft-output maximum a posteriori detector that provides sequences of a posteriori probabilities to the corresponding FEC decoders. It involves backward and forward recursions resulting in high complexity and processing delay. In this paper, we consider several suboptimum soft output disjoint multiuser detectors that involve only forward recursions and have reduced complexity and delay.

A parameterization of the cone around a given vector in N-dimensional vector space is derived. Vectors obtained by changing the parameter values can not escape the cone. The constraint is useful when it is known that an optimal vector of filter coefficients is close in direction with the given vector. Adaptive filtering algorithms that use the steepest descent method and stochastic gradient approximation are developed for the case of filter coefficients constrained in the cone. There is no need for monitoring the constraint since it is always satisfied by the construction. Two optimization criteria are explicitly considered: the least mean square and constant modulus. The cone constrained constant modulus algorithm (CMA) is applied to the problem of user detection in a synchronous direct sequence code division multiple access system. Its convergence is compared with the plain CMA and back projection CMA. Under severe conditions the cone constrained CMA is the only one which locks to the desired user.

We present an analytical framework which enables the MCSSA protocol defined and analyzed by Dastangoo, Vojcic and Daigle (see IEEE GLOBECOM '98) to support differentiated services (e.g., voice and data). The underlying principles to the MCSSA protocol consist of two elements. The first element concerns the users' random selection from a finite number of resources (i.e., channels or spread spectrum (SS) codes). The second is the simultaneous transmission of users' packets. These elements are realized through two stages of packet acquisition (Stage-I) and packet transmission (Stage-II). Therefore, a successful transmission depends not only upon the success in Stage-I, but also the success in Stage-II. To support multi-traffic classes requires several modifications to Stage-I of the MCSSA protocol. Naturally, voice traffic must be given preferential treatment over the data traffic, since the quality of service (QoS) of the former suffers more from excess access/queueing delays. In EMCSSA, we develop a priority mechanism in Stage-I which enables the voice users to contend, acquire, and reserve SS codes for the duration of their calls. Since the users' access to the resources in EMCSSA is distributed on a per-slot basis, each user's activity should be known to all. This is realized through a multi-level information feedback in Stage-I. We set the stage for both infinite population (Poisson/binomial model) and finite population (Markov model) cases. However, we only present the former. To assess the performance of the proposed protocol, we evaluate a commonly used performance metric: the average throughput.

In this paper, we present an analytical framework to investigate the dynamics of a family of medium access control (MAC) protocols which are governed by two distinct elements. The first element concerns the users' random selection from a finite number of resources (i.e., channels or spread spectrum (SS) codes). The second element is the simultaneous transmission of users' packets if the resources happen to be SS codes. These elements are realized through two stages of packet acquisition (Stage-I) and packet transmission (Stage-II). Our analytical model is properly constructed and adequately parameterized to reduce to the special cases: slotted ALOHA, multichannel slotted ALOHA, and CDMA. To assess the performance of the MCSSA protocol, we evaluate two commonly used performance metrics: average packet delay and throughput. To validate the Markov model (e.g., finite population case) developed in this paper, we also develop a Poisson model (e.g., infinite population case). Preliminary results indicate that, depending on the loading (e.g., high offered load level), the MCSSA protocol surpasses that of its CDMA counterpart.

Transmitter optimization, in addition to receiver optimization, contributes significantly to efficient interference suppression in multiple access and multipath channels. The system design is based on the joint optimization of the transmitter and the receiver in a synchronous multiuser channel characterized by multipath propagation. Joint optimization is represented by a linear transformation of the transmitted signals at the transmitter and a linear transformation of received signals at each receiving site that minimize the effect of multiple access and multipath interference. The minimum mean squared error between the true bit value and its estimate at the output of the receiver is taken as the cost function, subject to average and peak transmit power constraints. It is shown that joint transmitter-receiver optimization outperforms significantly either transmitter optimization or receiver-based techniques. The crucial assumption, in the case of multipath channels, is that the transmitter knows the multipath characteristics of all channels and that the channel dynamics are sufficiently slow so that multipath profiles remain essentially constant over a block of preceded bits. The practical applications can be found in indoor and cellular communications, satellite communications, or military communications where nonorthogonal signature waveforms are employed.

A synchronous multiuser system operating in an additive white Gaussian noise channel, with or without multipath fading, is considered. It is shown that when either a conventional single user receiver or the RAKE receiver is employed, both multiple access and intersymbol interference can be eliminated by means of a suitable transmitter precoding scheme. Transmitter precoding represents a linear transformation of transmitted signals, such that the mean squared errors at all receivers are minimized. Precoding, with both conventional single user receiver and with the RAKE receiver, results in near-far resistant performance and outperforms considerably the respective schemes without precoding. The crucial assumption, in the multipath case, is that the transmitter knows the multipath characteristics of all channels and that channel dynamics are sufficiently slow so that multipath profiles remain essentially constant over the block of precoded bits.

In recent years, a great deal of research has been performed on methods of alleviating the performance degradations suffered by code division multiple access (CDMA) systems due to multiple access interference. The majority of work thus far has focused on the coherent case, spurred by the development of an optimal detector for asynchronous binary phase shift keying (BPSK) by Verdu. Much less work has been performed on noncoherent detection. There are, however, several reasons why noncoherent detection may be preferred for some applications. For instance, there may be difficulty tracking carrier phase over some channels due to oscillator phase instability, movement of the transmitter and receiver in a mobile environment, and multipath fading. In this paper, a multistage detector for noncoherent CDMA using M-ary orthogonal signals is proposed and analyzed. Using a decorrelating detector as its first stage, the detector iteratively estimates the multiple access interference affecting the desired signal, subtracts the estimated interference, and forms symbol estimates for each of K users. Numerical examples are provided to contrast the bit error rate performance of the decorrelating detector with previous detectors for the same signalling scheme.

We are interested in determining the feasibility of having multiple service providers share the same spectrum in transponder-type. Satellite-based, code-division multiple-access mobile communications. To accomplish this, we compare the total capacity when n systems are operational to that when only a single system uses the frequency band. This comparison is made for both the uplink and the downlink under one of two possible constraints, either a maximum power flux density (PFD) limit per system, or an aggregate PFD limitation. While there are circumstances when the total capacity of multiple systems can achieve a greater value than that of a single service provider, conditions depend on the degree of shadowing, and whether or not the PFD limit is on a per system basis or on an aggregate basis from all systems.>

A DS CDMA satellite uplink with power control error is considered. CDMA system performance is normally interference limited, and individual users typically experience the average performance over all users. This is particularly true in a Gaussian, additive white noise channel, with many users, sufficient processing gain, and effective power control. In this case, the other users appear as Gaussian noise and each user experiences the same environment. Under fading channel conditions however, the average behavior notion is flawed, since the fade conditions for each user are different, even if the fading statistics are identical. An alternative measure of performance under these conditions is the probability of outage. For coded signals, the fading must be averaged over all the symbols in the codeword. For properly interleaved signals and i.i.d. Rayleigh fade statistics, the average must be performed over a chi-squared variate. The probability of outage, defined as the probability that the decoded bit error rate exceeds a specified threshold, is monotonic with the probability that the chi-squared variate is less than a specified threshold. The authors derive upper and lower bounds on the outage probability for soft decision decoding.

The analysis of both the performance and capacity of direct sequence CDMA in terrestrial cellular systems has been addressed in the technical literature. It has been suggested that CDMA be used as a multiple access method for satellite systems as well, in particular for multispot beam low Earth orbit satellites (LEOS). One is tempted to argue that since CDMA works well on terrestrial links, it will nominally work as well on satellite links. However, because there are fundamental differences in the characteristics of the two channels, such as larger time delays from the mobile to the base station and smaller multipath delay spreads on the satellite channels, the performance of CDMA on satellite links cannot always be accurately predicted from its performance on terrestrial channels. In the paper, the authors analytically derive the performance of a CDMA system which operates over a low Earth orbiting satellite channel. They incorporate such effects as imperfect power control and dual-order diversity to obtain the average probability of error of a single user.>

Consideration is given to the performance of a binary phase-shift-keyed (BPSK) direct-sequence spread-spectrum (DS/SS) communication system in a fading dispersive channel with the direct or specular component of the signal present. Two types of multipath intensity profiles are analyzed. Large values of the multipath spread (larger than a symbol duration) are considered, and a useful approximation to the symbol error probability, based on random spreading sequences, is derived. The effect of pulsed jamming on the performance is analyzed. Both the uncoded and the coded case are considered. In the coded case, hard- and soft-decision decoding metrics, with and without jammer-state information (JSI), are analyzed and discussed.>

The performance of a binary phase-shift-keyed direct-sequence spread-spectrum communication system in a fading dispersive channel with jamming is examined. The channel is modeled as a wide-sense-stationary uncorrelated scattering channel, dispersive only in time. The channel output consists of a strong stable component of the signal plus a continuum of dispersive components and is referred to as a frequency-selective Rician fading channel. To obtain specific results, the multipath intensity profile is modeled by triangular function. Both the uncoded and coded performances are analyzed. The multipath spread is chosen to limit intersymbol interference to K adjacent coded symbols in the coded case and to only adjacent data bits in the uncoded case. No specific interleaver design is considered. A closed-form expression for the probability of error is obtained, assuming random spreading sequences. Numerical results are presented for values of the system and channel parameters. It is shown that a jammer can optimize its strategy if the probability of error is not too high.>