Stochastic gradient algorithms for design of minimum error-rate linear dispersion codes in MIMO wireless systems

Linear dispersion (LD) codes are a good candidate for high-data-rate multiple-input multiple-ouput (MIMO) signaling. Traditionally LD codes were designed by maximizing the average mutual information, which cannot guarantee good error performance. This paper presents a new design scheme for LD codes that directly minimizes the block error rate (BLER) in MIMO channels with arbitrary fading statistics and various detection algorithms. For MIMO systems employing LD codes, the error rate does not admit an explicit form. Therefore, we cannot use deterministic optimization methods to design the minimum-error-rate LD codes. In this paper, we propose a simulation-based optimization methodology for the design of LD codes through stochastic approximation and simulation-based gradient estimation. The gradient estimation is done using the score function method originally developed in the discrete-event-system community. The proposed method can be applied to design the minimum-error-rate LD codes for a variety of detector structures including the maximum-likelihood (ML) detector and several suboptimal detectors. It can also design optimal codes under arbitrary fading channel statistics; in particular, it can take into account the knowledge of spatial fading correlation at the transmitter and receiver ends. Simulation results show that codes generated by the proposed new design paradigm generally outperform the codes designed based on algebraic number theory.     

Multiuser square-law detection with applications to fiber optic communications

Optimal and suboptimal multiuser noncoherent detection for square-law receivers is studied in this paper. Great potential is found for the multiuser square-law detector when compared with the conventional single-user square-law detector. We study the optimal detector and two detectors with simpler structures: the asymptotically optimal detector and the pairwise linear detector. The saddle-point approximation method is used to study the error performance and its asymptotic behavior as noise reduces. Due to difficulty in optimizing the detectors at an arbitrary noise level, we propose to use the asymptotic efficiency for detector optimization. For a low-error-rate system like a fiber optic communication system, the asymptotic efficiency is found to be an efficient way to design detectors. Numerical results show that the asymptotically optimized detectors perform as well as the optimal detector, even for nonzero noise levels of practical interest. Despite their exponential complexity, these detectors are applicable to wavelength-division multiplexed systems in which only a few neighboring channels produce strong interference.     

Penalties of Sample-and-Sum and Weighted Partial Decision Detectors in Gaussian Noise

The deterioration in performance, measured in the probability of error sense, of sample-and-sum and weighted partial decision detectors are analyzed for arbitrary signal-to-noise ratios. These suboptimal detectors have more modest computational requirements than the optimal digital matched filter making them amenable to simple digital implementations. The effects on the penalties of the signaling waveform employed, the number of samples processed, and the signal-to-noise ratio are considered in detail. Included are the penalties for the optimum weighted partial decision detector. The optimum weighted partial decision detector is the optimum detector, in the minimum probability of error sense, for hard-limited samples. The penalty of the optimum weighted partial decision detector relative to the digital matched filter detector represents the fundamental loss in signal detectability due to hard-limiting in a sampled system.     

Design of Spherical Lattice Space–Time Codes

In this paper, we propose a systematic procedure for designing spherical lattice (space–time) codes. By employing stochastic optimization techniques we design lattice codes which are well matched to the fading statistics as well as to the decoder used at the receiver. The decoders we consider here include the optimal albeit of highest decoding complexity maximum-likelihood (ML) decoder, the suboptimal lattice decoders, as well as the suboptimal lattice-reduction-aided (LRA) decoders having the lowest decoding complexity. For each decoder, our design methodology can be tailored to obtain low error-rate lattice codes for arbitrary fading statistics and signal-to-noise ratios (SNRs) of interest. Further, we obtain fundamental lower bounds on the error probabilities yielded by lattice and LRA decoders and characterize their asymptotic behavior.      

Analysis of Differential Orthogonal Space–Time Block Codes Over Semi-Identical MIMO Fading Channels

We study the performance of differential orthogonal space-time block codes (OSTBC) over independent and semi-identically distributed block Rayleigh fading channels. In this semiidentical fading model, the channel gains from different transmit antennas to a common receive antenna are identically distributed, but the gains associated with different receive antennas are nonidentically distributed. Arbitrary fluctuation rates of the fading processes from one transmission block to another are considered. We first derive the optimal symbol-by-symbol differential detector, and show that the conventional differential detector is suboptimal. We then derive expressions of exact bit-error probabilities (BEPs) for both the optimal and suboptimal detectors. The results are applicable for any number of receive antennas, and any number of transmit antennas for which OSTBCs exist. For two transmit antennas, explicit and closed-form BEP expressions are obtained. For an arbitrary number of transmit antennas, a Chernoff bound on the BEP for the optimal detector is also derived. Our results show that the semi-identical channel statistics degrade the error performance of differential OSTBC, compared with the identical case. Also, the proposed optimal detector substantially outperforms the conventional detector when the channel fluctuates rapidly. But in near-static fading channels, the two detectors have similar performances     

Optimal and mismatched detection of QAM signals in fast fading channels with imperfect channel estimation

In this paper, we derive an optimal detector for pilot-assisted transmission in Rayleigh fading channels with imperfect channel estimation. The classical approach is based on obtaining channel estimates and treating them as perfect in a minimum distance detector (this is called mismatched detector). The optimal detector jointly processes the received pilot and data symbols to recover the data. The optimal detector is specified for fast frequency-flat fading channels.We consider spline approximation of the channel gain time variations and compare the detection performance of different mismatched detectors with the optimal one. Further, we investigate the detection performance of an iterative receiver in a system transmitting turbo-encoded data, where a channel estimator provides either maximum likelihood estimates, minimum mean square error (MMSE) estimates or statistics for the optimal detector. Simulation results show that the optimal detector outperforms the mismatched detectors. However, the improvement in the detection performance compared to the mismatched detector with the MMSE channel estimates is modest.     

Channel knowledge and optimal performance for two-wave Rayleighfading channels

The optimal detector structures and error probability performances for the two-wave Rayleigh fading channel with known delay are compared for different levels of channel knowledge. A number of different detectors are examined, for equal energy signals having equal complex autocorrelation magnitudes. Envelope orthogonal frequency-shift keying and variants of chirp signals are considered, so that the complex cross correlation of the signals is unconstrained. The performance of the optimal detector, when the fading in neither wave is tracked, is obtained. Two other, suboptimal, quadratic detectors are also considered for this case. Optimal detection, when the fading in only one of the waves is tracked, while a statistical knowledge of the other wave is available, is examined. The optimal performance that can be achieved by a time-varying matched-filter detector that makes use of complete knowledge of the channel fading in both waves is also determined. These detectors, for all the different levels of channel information considered, are studied in a unified framework, the probability of error being expressed as the probability that a quadratic form in Gaussian random variables is less than zero. It is found that the power gain that can be derived from partial or complete tracking is small. All these detectors exhibit a diversity-like effect for all nonzero values of the delay and for most values of the signal parameters. Signals with larger dispersion factors, such as chirp signals and variants, perform well on the channel, enhancing the diversity effect, even at small delays     

On the detectability of weak signals (Corresp.)

Approximate performance results are obtained for optimal detection of weak discrete time signals. The analysis parallels the "locally optimum" approach for detector design and is based on a small-signal Taylor series approximation to the Chernoff error probability bounds.     

Multiuser detection in asynchronous CDMA frequency-selective fading channels

Multipath fading severely limits the performances of conventional code division multiple-access (CDMA) systems. Since every signal passes through an independent frequency-selective fading channel, even modest cross-correlations among signature sequences may induce severe near-far effects in a central multiuser receiver. This paper presents a systematic approach to the detection problem in CDMA frequency-selective fading channels and proposes a low complexity linear multiuser receiver, which eliminates fading induced near-far problem.We initially analyze an optimal multiuser detector, consisting of a bank of RAKE filters followed by a dynamic programming algorithm and evaluate its performance through error probability bounds. The concepts of error sequence decomposition and asymptotic multiuser efficiency, used to characterize the optimal receiver performance, are extended to multipath fading channels.The complexity of the optimal detector motivates the work on a near-far resistant, low complexity decorrelating multiuser detector, which exploits multipath diversity by using a multipath decorrelating filter followed by maximal-ratio combining. Analytic expressions for error probability and asymptotic multiuser efficiency of the suboptimal receiver are derived that include the effects of multipath fading, multiple-access interference and signature sequences correlation on the receiver's performance.The results indicate that multiuser detectors not only alleviate the near-far problem but approach single-user RAKE performance, while preserving the multipath diversity gain. In interference-limited scenarios multiuser receivers significantly outperform the RAKE receiver.This paper was presented in part at the Twenty-Sixth Annual Conference on Information Sciences and Systems, Princeton, NJ, March 1992 and MILCOM'92, San Diego, CA, October 1992. This work was performed while author was with the Department of Electrical and Computer Engineering, Northeastern University, Boston, USA.     

Group detection for synchronous Gaussian code-divisionmultiple-access channels

The concept of group detection Is introduced to address the design of suboptimum multiuser detectors for code-division multiple-access (CDMA) channels. A group detection scheme consists of a bank of P group detectors, one each for detecting the information symbols of users in each group of a P group partition of the K simultaneously transmitting users. In a parallel group detection scheme, these group detectors operate independently, whereas in a sequential scheme, each group detector. Uses the decisions of the previous group detectors to successively cancel the interference from those users. Group detectors based on the generalized likelihood ratio test (GLRT) are obtained for the synchronous Gaussian CDMA channel. The complexity of these detectors is exponential in the group size, whereas that of the optimum detector is exponential in K. Since the partition of users is a design parameter, group sizes can be chosen to satisfy a wide range of complexity constraints. A key performance result is that the GLRT group detectors are optimally group near-far resistant. Furthermore, upper and lower bounds on the asymptotic efficiency of the sequential group detectors are derived. These bounds reveal that the sequential group detectors can, under certain conditions, perform as well as GLRT group detectors of much larger group sizes. Group detection provides a unifying approach to multiuser detection. When the users are partitioned into K single-user groups, the GLRT, a modified form of GLRT, and the sequential group detectors reduce to previously proposed suboptimal detectors; namely, the decorrelator, the two-stage detector, and the decorrelating decision-feedback detector, respectively. For the other nontrivial partitions, the group detectors are new and have a performance that is commensurate with their complexity     

Detection Performance Considerations for Direct-Sequence and Time-Hopping LPI Waveforms

The implementation and performance of wide-band detectors for direct-sequence and time-hopping spread-spectrum waveforms in the presence of additive white Gaussian noise are considered in this paper. Of interest here is the performance penalty incurred when going from optimal to suboptimal detector structures. In both cases, performance is quantified by appropriately defined distance measures and is ultimately compared to that of the simplest hypothesis-discriminating device, namely, the energy detector (radiometer).     

A Class of Soft Decision Error Detectors for the Gaussian Channel

The main idea of this concise paper is to use symbol reliability information for improving the performance of error detectors. A class of soft decision error detectors (SDED's) is presented, where low weight error patterns are selectively corrected. The reliability numbers govern which error pattern to correct. Asymptotic performance and bounds are derived for the Gaussian channel. It is shown that singleerror correcting SDED's without thresholds have an exponentially lower repeat request probability than the hard decision error detector, but the same exponent in the probability of undetected errors. Detectors with thresholds are also considered. In this case it is possible to contruct an error detector with both better probability of undetected errors and better probability of repeat request than for the hard decision error detector (HDED). Bounds and computer simulations are presented and SDED's are compared to HDED's.     

Performance evaluation of differential and discriminator detection of continuous phase modulation

Recently, bandwidth efficient constant-amplitude digital modulation schemes have also been shown to be power efficient when coherent detection is used. Partial-response continuous phase modulation (CPM) schemes are within this class. In some applications noncoherent detection is preferred. The performance of CPM systems is analyzed for differential and discriminator detection. An additive white Gaussian channel is assumed. The detectors make symbol-by-symbol decisions. The considered schemes are M-ary with an arbitrary modulation index and pulse shaping over several symbol intervals. The performance is analyzed by means of error probability expressions. The IF filter for the detectors is optimized within a special class of filters to give good performance. The differential detector is also analyzed on a Rayleigh fading channel. The fading is assumed to be slow. The IF filter is also optimized on this channel. Simulated error probabilities for discriminator detection with a Viterbi detector are also presented both for the Gaussian and the Rayleigh fading channel. The discriminator detector making symbol-by-symbol decisions is simulated on the Rayleigh fading channel. It is shown that partial-response CPM schemes with good performance can also be obtained with noncoherent detectors.     

Aspects on single symbol signaling on the frequency flat Rayleighfading channel

The optimal single symbol detector on the Rayleigh fading channel computes a functional quadratic form in the time continuous received signal. A drawback is that closed-form solutions of integral equations based on the channel statistics are required. This makes simplified discrete receivers attractive. A class of suboptimal receivers that transforms the received random process to a set of discrete observables is derived. The set of observables constitutes a random vector in a finite dimensional receiver signal space. Given this vector, the maximum likelihood detector computes a quadratic form in the received vector. The discretization implies a loss of information, therefore such a detector is not, in general, optimal given the received time continuous signal. The purpose is, however, to achieve close to optimal performance when the number of observables becomes large. This class of detectors is analyzed using exact error probability calculations, which reveal several interesting properties. The length of the observation interval and the number of discrete observables have significant influences on the error probability when the time variations of the fading process are rapid compared with the symbol duration. By increasing the number of observables, the error floor is lowered, and the implicit diversity order is increased. This implicit diversity arises as soon as more than one observable per symbol interval is used and is a consequence of the information bearing signal being a random process. Matched filter receivers use few discrete observables per symbol interval, and thus suffer from high error floors and low implicit diversity orders on fast fading channels. The error probability is highly dependent on the shapes and durations of the modulator waveforms. For instance, pulses of long duration give lower error probabilities than shorter pulses, and for a certain type of orthogonal waveforms there is no error floor     

Transmit Antenna and User Selection for Multiuser Spatial Multiplexing with Maximum Likelihood Receiver

In this paper, we consider a transmit antenna and user selection for maximum-likelihood (ML) detector in multiuser spatial multiplexing systems. The conventional algorithm that maximizes the minimum stream SNR is not optimal in terms of the error probability, because it is tailored for linear detectors. We propose a simple transmit antenna and user selection scheme for the ML detector based on maximizing the minimum distance. Numerical results show that the proposed scheme provides enhanced performance compared with the conventional algorithms in terms of the error probability of ML receiver.     

Further Studies of Soft Decision Error Detectors for the Gaussian Channel

This correspondence is an extension of the work in [1] and [2]. We present two new classes of error detectors which utilize symbol reliability information (i.e., soft decision error detectors). First, a simple class of selectively single error correcting error detectors is discussed where the syndrome information is used and where just one single reliability number is compared to a threshold value. These error detectors are simpler than those in [1] and [2], but less powerful. Second, a class of soft decision error detectors with low probability of undetected errors is considered. The reliability information is used to reduce the probability of undetected errors without increasing the repeat request probability (compared to the hard decision error detector). Exponential properties for large signal-to-noise ratios and bounds are calculated for both classes of error detectors. Computer simulations are also presented.     

Soft information transfer for sequence detection with concatenatedreceivers

Concatenated encoders are found in many digital communication systems. A concatenated receiver for such a system consists of a chain of independently working estimators ended with an outer detector. Each stage of the receiver chain corresponds to a certain code in the system. Maximum likelihood sequence detection with a concatenated receiver requires soft information consisting of symbol aposteriori probabilities to be transferred between the receiver stages. In this paper the optimal inner estimator is derived and further simplified into a new asymptotically optimal one by imposing a tree structure on the inner code. By mapping several suboptimal inner estimators onto the tree structure and noting which approximations they make, a theoretical performance ranking is obtained. The asymptotical performance of the optimal concatenated detector at high signal-to-noise ratios (SNRs) is investigated. The decision variables underlying the pairwise error event probability are shown to be asymptotically Gaussian and consequently a free Euclidean distance can be found for the optimal concatenated detector     

Error performance of multiple-symbol differential detection of PSKsignals transmitted over correlated Rayleigh fading channels

The error performance of multiple-symbol differential detection of uncoded QPSK signals transmitted over correlated Rayleigh fading channels is studied. The optimal detector is presented, along with an exact expression for the corresponding pairwise error event probability. It is shown that multiple-symbol differential detection is a very effective strategy for eliminating the irreducible error floor associated with a conventional differential detector. In all of the cases investigated, a detector with an observation interval as small as two symbols is sufficient for this purpose. It is also found that the error performance of a multiple-symbol differential detector is not sensitive to the mismatch between the decoding metric and the channel fading statistics     

Robust detection in digital communications

This paper deals with the problem of communicating through unspecified noise. Detectors, robust against variations in the probability density function of the noise, are developed and discussed. The paper covers three issues. First, the relation between a distance measuring receiver and a correlating receiver in a general case is shown. Second, a theoretical method for the computation of an upper limit for the probability of symbol error is presented. This computation fits into the ordinary framework for computation of the error probability by changing the inverted noise density 2/N₀ to efficacy, ϵ. Efficacy is defined in the paper. Third, detectors based on M-, i.e., maximum likelihood type, and R-, i.e., rank, statistics are tested and compared for GMSK and π/4-shifted DQPSK. From numerical comparisons of the upper bounds and their simulated estimates for robust detectors, it is concluded that the loss in Gaussian noise is very small compared to the optimum quadratic detector. The gain, compared to a nonrobust receiver optimized to Gaussian noise, is around 0.5 to 2 dB for large SNR and around 2 to 4 dB for low SNR in impulsive noise. This offers new methods of significantly improving communication when the noise is unknown     

Simulation Results for the Decision-Directed MAP Receiver for M-ary Signals in Mulitiplicative and Additive Gaussian Noise

Simulation results are presented for the error-rate performance of the recursive digital maximum a posteriori probability (MAP) detector for knownM-ary signals in multiplicative and additive Gaussian noise. The structure of the digital simulation of the optimum detector is generally described, with specific results obtained for a quaternary signal and 2500 digit per second transmission rate. The simulation is focused on the aeronautical multipath communication problem. Plots of detection error rate versus additive signal-to-noise ratio are given, with the power ratio of multiplicative process to desired signal as a parameter. Results are presented for the cases where the detector has perfect knowledge of the first- and second-order statistics of the multiplicative and additive processes and also where these statistics are estimated in near real time. For comparison, the error rates of conventional coherent and noncoherent digital MAP detectors are also obtained. It is shown that with nonzero multiplicative noise, the error rates of the conventional detectors saturate at a level that is irreducible for increasing additive signal-to-noise ratio. The error rate of the optimum detector having perfect statistical knowledge continues to decrease rapidly with increasing additive signal-to-noise ratio. In the absence of multiplicative noise, the conventional coherent detector and the optimum detector are shown to exhibit identical performance. Suboptimum detectors, having less than perfect statistical knowledge, yield error rates bounded below by the optimum detector rates and bounded above by the conventional detector rates.