A composite source, consisting of multiple subsources and a memoryless switch, outputs one symbol at a time from the subsource selected by the switch. If some data should be encoded more accurately than other data from an information source, the composite source model is suitable because in this model different distortion constraints can be put on the subsources. In this context, we propose subsource-dependent fidelity criteria for composite sources and use them to formulate a rate-distortion problem. We solve the problem and obtain a single-letter expression for the rate-distortion function. Further rate-distortion analysis characterizes the performance of classify-then-compress (CTC) coding, which is frequently used in practice when subsource-dependent fidelity criteria are considered. Our analysis shows that CTC coding generally has performance loss relative to optimal coding, even if the classification is perfect. We also identify the cause of the performance loss, that is, class labels have to be reproduced in CTC coding. Last but not least, we show that the performance loss is negligible for asymptotically small distortion if CTC coding is appropriately designed and some mild conditions are satisfied.

To deal with the multimodality, nonlinearity, and nonconvexity of the maximum likelihood (ML) cost function when solving the localization problem with received signal strength (RSS) in wireless sensor networks, we propose a localization approach based on the differential evolution algorithm, opposition-based learning, and adaptive redirection in a unified way. In the proposed approach, we need neither to approximate the ML cost function nor to provide a good initial point while most conventional approaches, including those based on the semidefinite programming, second-order cone programming, and unscented transformation. The performance and computational complexity of the proposed approach are analyzed and compared with those of other approaches for three practical scenarios. Simulation results confirm that the proposed approach performs better than other approaches with relatively low computational complexity. Note to Practitioners-This article was motivated by the problem of localization in wireless sensor networks that can be found in different areas, such as intelligent farming, health care monitoring, and environmental monitoring. In this article, a technique is proposed, which exploits the RSS measurements to estimate the position of a target node. This technique is quite popular due to its simplicity and low cost. Most of the techniques based on the ML formulation approximate the ML cost function due to its high nonlinearity and nonconvexity. This article suggests a new approach using differential evolution and opposition-based learning a redirection, which does not require approximating the ML cost function to find the location of a target. In addition, in this article, we analyze the sensor nodes' placement and its effect on localization accuracy. Numerical results in three practical scenarios are provided to demonstrate the effectiveness of the proposed approach and its superiority compared with state-of-the-art algorithms. In future research, we will address a more realistic scenario with multiple targets.

Taking the positions of pico-cell base stations (PBSs) into consideration, a scheme of cell range expansion (CRE) for maximum sum rate is addressed in heterogeneous multi-input multi-output multi-user wireless networks. The optimal CRE bias obtained numerically by the proposed CRE scheme with intercell interference coordination (ICIC) allows us to maximize the sum rate while successfully maintaining the load balance between the macro-cell base station and PBSs. Numerical results confirm that the proposed CRE scheme with ICIC can provide a higher sum rate than conventional schemes while maintaining balanced load.

We consider a simultaneous wireless transfer of information and power in a two-way relaying network, where a decode-and-forward protocol is employed for data exchange between two source devices at different rates. The data exchange is accomplished in two time phases (TPs) with possibly asymmetric TP ratios, where the relay is powered by the source devices in the first phase through simultaneous wireless information and power transfer (SWIPT) either by power splitting (PS) or by time switching (TS). For the network, resource allocation of the PS and TP ratios for the PS-based SWIPT and that of the TS and TP ratios for the TS-based SWIPT are studied to minimize the system outage probability. We first solve the joint optimization problem of each case with the instantaneous channel state information and then propose suboptimal schemes utilizing one or two values for the TP ratio to lower the implementation complexity. The results from analysis and simulation show that the proposed schemes outperform the benchmark scheme with equal resource allocation and that the gain gets more prominent as the rates of the data exchanged or the relay locations become more asymmetric.

A class of low-complexity multi-antenna relaying schemes is proposed for orthogonal frequency division multiplexing based one-way and two-way relay networks. In the proposed schemes, two single-antenna terminal nodes complete the data transfer in two time phases via a multi-antenna amplify-and-forward relay without channel state information, discrete Fourier transform (DFT), and inverse DFT. The relay signal processing consists of receive combining, power scaling, and transmit diversity, where instantaneous time domain power scaling is proposed for power scaling and power-based selection combining and cyclic delay combining are proposed for receive combining, to leverage the performance only with time domain operations. An approximation to the outage probability is also derived to predict the coded performance behavior under practical channel models. It is shown by evaluating the outage probability and coded bit error rate that the proposed schemes outperform the conventional schemes of comparable complexity.

We address detection schemes of spectrum sensing for cognitive radio with multiple receive antennas operating over a wideband channel composed of a multitude of subbands. By taking the observations in all subbands into consideration in the likelihood functions for sensing a subband, the test statistics of the proposed schemes are functions of the sample covariance matrix in the subband under consideration and that in the subband exhibiting the lowest power spectral density. The false alarm and detection probabilities of the proposed schemes are analyzed theoretically and confirmed via simulations when the numbers of observations are the same for all the subbands. It is shown through computer simulations that the proposed schemes can provide considerable performance gains over conventional schemes for wideband spectrum sensing when the observations are spatially correlated and temporally independent/dependent.

We consider two-way relay (TWR) networks consisting of two communicating nodes and multiple relays built upon orthogonal frequency-division multiplexing signaling. For low-cost implementation of relays, two pragmatic TWR protocols are proposed based on analog network coding (ANC). The proposed two-phase (2P)- and three-phase (3P)-ANC schemes employ time-domain instantaneous power scaling and relay selection to improve the outage performance without equipping discrete Fourier transform (DFT) and inverse DFT functions at the relay. In particular, we devise a relay power allocation factor and a resultant relay selection metric for the proposed 3P-ANC scheme, which can be implemented at low complexity. For the proposed schemes, the outage probability and achievable diversity order are also analyzed, which reveals a larger diversity order of the 3P-ANC scheme than the 2P-ANC scheme. The results from analysis and simulation show that the proposed schemes outperform the conventional 2P-ANC schemes employing time-domain average power scaling and distributed space-time block coding or frequency-domain power scaling and relay selection.

A detection scheme of spectrum sensing is discussed for cognitive radio with multiple receive antennas operating over a wideband channel composed of a multitude of sub bands. By taking the observations in all sub bands into consideration in the likelihood ratio test for sensing a sub band, the proposed scheme can provide better performance than other conventional schemes.

We consider a relay network that delivers a Gaussian source by employing successive refinement source coding and superposition coding of layers at the source node and successive decoding at the relay and destination nodes. For the network, making use of the decoding results at the relay and destination nodes of the first transmission, an efficient relaying strategy of layers is proposed to minimize the expected distortion (ED) when only the average channel state information is available at the source node. Three types of the proposed scheme, defined as Prop-DF, using decode-and-forward signals, Prop-AF, using amplify-and-forward signals, and Prop-MF, using mixed-forward signals, are addressed and analyzed in terms of the outage probability and distortion exponent. Unlike other studies, we have also taken the relay location into account in deriving the distortion exponent showing the high SNR behavior of the ED. The results show that the proposed scheme increases the distortion exponent up to twice that of the conventional relaying schemes when the relay is close to the source node, and that Prop-MF provides the best performance for most relay locations.

Owing to the singularity of the within-class scatter, linear discriminant analysis (LDA) becomes ill-posed for small sample size (SSS) problems. Null-space-based LDA (NLDA), which is an extension of LDA, provides good discriminant performances for SSS problems. Yet, as the original scheme for the feature extractor (FE) of NLDA suffers from a complexity burden, a few modified schemes have since been proposed for complexity reduction. In this brief, by transforming the problem of finding the FE of NLDA into a linear equation problem, a novel scheme is derived, offering a further reduction of the complexity.

To reduce the influence of interference on the OFDM-based cognitive radio overlaid with narrowband primary users, we consider a scheme called the variable subband nulling (VSN), in which a VSN mode is selected adaptively according to the result of spectrum sensing for the frames following the sensing period. For reliable detection of the VSN mode embedded in the preamble, we consider two types of metrics. The performances of the metrics are analyzed in Gaussian channels with/without narrowband interference (NBI), and are confirmed by simulation results. In fading channels with/without NBI, the detection performance is examined via simulation results. The analysis and simulation results have confirmed that, if the VSN mode is selected appropriately at the transmitter, the proposed metrics provide us with a reliable detection performance even when the NBI is strong.

A class of recurrent neural networks is proposed and proven to be capable of identifying any discrete-time dynamical system. The application of the proposed network is addressed in the encoding, identification, and extraction of finite state automata (FSAs). Simulation results show that the identification of FSAs using the proposed network, trained by the hybrid greedy simulated annealing with a modified cost function in the training stage, generally exhibits better performance than the conventional identification procedures.

This paper presents a cross-layer design of adaptive modulation and coding (AMC) with truncated hybrid automatic repeat request (HARQ) for one-to-many multicast transmission, in order to increase the spectral efficiency under the quality-of-service (QoS) constraints. For adaptive transmission over the channel common to the multicast group users, the AMC mode is chosen with the minimum SNR among the users to guarantee the target performance of all users. In the meanwhile, the minimum SNR required to support an AMC mode is aggressively designed by allowing retransmission with HARQ schemes. For the proposed design, we derive the average packet error rate, average number of transmission, and spectral efficiency and provide the performance numerically obtained. Numerical results show that the cross-layer design for multicast provides a significant performance gain at a small number of retransmissions as in unicast. In particular, it is observed that AMC design with HARQ is more beneficial in the low SNR region where multicast performs worse than unicast.

In this paper, we propose a near maximum likelihood (ML) scheme for the decoding of multiple input multiple output systems. Based on the metric-first search method and by employing Schnorr-Euchner enumeration and branch length thresholds, the proposed scheme provides a higher efficiency than other conventional near ML decoding schemes. From simulation results, it is confirmed that the proposed scheme has lower computational complexity than other near ML decoders while maintaining the bit error rate very close to the ML performance. The proposed scheme in addition possesses the capability of allowing flexible tradeoffs between the computational complexity and BER performance.

In this paper, we propose a novel ML decoding scheme based on the combination of depth- and breadth-first search methods on a partitioned tree for multiple input multiple output (MIMO) systems. The proposed scheme first partitions the searching tree into several stages, each of which is then searched by a depth- or breadth-first search method, maximally exploiting the advantages of both the depth- and breadth-first search methods. Numerical results indicate that, when the depth- and breadth-first search algorithms are adopted appropriately, the proposed scheme exhibits substantially lower computational complexity than conventional ML decoders while maintaining the ML bit error performance.

For the CR systems sharing the spectrum with narrowband primary devices, this paper presents an OFDM system supporting the fractional bandwidth (FBW) mode to avoid the interference to or from the primary devices. For the system, the preamble and FBW mode detection algorithm are provided to obtain the FBW mode information during the initial synchronization. The simulation results reveal that the FBW mode information can be obtained at the SNR region lower than Es/N0 = 0 dB. In addition, the FBW mode transmission and detection do not incur any deterioration of the synchronization performance.

In most previously reported studies on locally optimum detection of signals, independent observations have been assumed in various noise environments. The use of an independent observation model may cause a considerable performance degradation in detection applications of modern high data-rate communication systems exhibiting dependence among interference components. In this paper, the detection of weak known and random signals is addressed in observations corrupted by multiplicative and first-order Markov additive noises. The asymptotic and finite sample-size performance of several detectors are obtained and compared: it is confirmed that the dependence of noise components need to be taken into account to maintain detection performance appropriately.

The performance characteristics of the cell averaging (CA), greatest of (GO), and smallest of (SO) constant false alarm rate (CFAR) processors in nonhomogeneous environment are analyzed and compared when receiving antenna diversity is employed in the pseudonoise (PN) code acquisition of direct- sequence code division multiple access (DS/CDMA) systems. From the simulation results, the GO CFAR processor is observed to exhibit the best performance in nonhomogeneous environment, with the CA CFAR processor performing the second best.

The use of an independent sample model may cause a considerable performance degradation in detection applications of modern high data rate communication systems exhibiting dependence among interference components. In this paper, we address the detection of weak known signals in observations corrupted by multiplicative and first-order Markov additive noise. The asymptotic performance of several detectors are obtained and compared, confirming that the dependence among interference components need to be taken into account to maintain performance accordingly.

The performance of weak signal detectors is addressed in additive noise described by the first order moving average (FOMA) of an impulsive process. The decision regions of the maximum likelihood (ML) and suboptimum ML (S-ML) detectors are derived in the FOMA model. The ML and S-ML detectors are employed in the antipodal signaling system, and compared in terms of the bit-error-rate performance in impulsive environment. Numerical results show that the S-ML detector, despite its reduced complexity and simpler structure, exhibits practically the same performance as the ML detector. It is also shown that the performance gap between detectors for FOMA and independent and identically distributed noise becomes larger as the degree of noise impulsiveness increases.

In this correspondence, we address the derivation of joint distributions and correlation coefficients for seven pairs of statistics used commonly in a number of signal detection schemes. The upper and lower bounds of the correlation coefficients are obtained, and relationships among the correlation coefficients are derived. Explicit values of the correlation coefficients evaluated for some specific noise distributions are shown for easy reference.

In this letter, we propose a new detector for code acquisition systems in non-Gaussian noise channels. Modeling the acquisition problem as a hypothesis testing problem, a detector is derived for non-Gaussian, symmetric /spl alpha/-stable noise, based on the locally optimum detection technique. Numerical results show that the proposed detector can offer robustness and substantial performance improvement over the conventional schemes in non-Gaussian channels.

We investigate optimal and suboptimal decision rules for parallel code acquisition in chip-asynchronous direct-sequence spread-spectrum (DS/SS) systems. The conventional decision rule for parallel acquisition is to choose the largest correlator output of a receiver. However, such a scheme is optimum only for chip-synchronous models. In this letter, an optimal decision rule is derived based on the maximum-likelihood criterion for chip-asynchronous models. A simpler suboptimal decision rule is also discussed. The performance of the optimum and suboptimum decision rules is compared to that of the conventional decision rule. Numerical results show, that for chip-asynchronous models, both the optimal and suboptimal decision rules outperform the conventional decision rule.

This paper presents a turbo coded multicarrier direct sequence code division multiple access (DS/CDMA) system, where the outputs of a turbo encoder are repetition coded at multiple rates and transmitted in parallel over a number of subchannels. A performance bound useful in the so-called error floor region is obtained for Rayleigh fading channels when different diversity orders are given to each turbo code symbol. Simulation results are also provided for the low signal-to-noise ratio (SNR) region where the bound is not applicable. It is observed that the error floor can be lowered, with some performance loss in the low SNR region, by applying nonuniform repetition coding to the turbo code symbols.

In this paper, a new class of polyphase sequences for CDMA systems and its generation method are suggested, and the correlation properties of the sequence are investigated. The performance of the sequence in quasi-synchronous CDMA (QS-CDMA) systems is investigated under a frequency-selective, time-nonselective, slow Nakagami fading channel with additive white Gaussian noise (AWGN). It is shown that the performance of the QS-CDMA systems using the suggested sequence is independent of the number of users and is much better than that using the PN sequence.

We consider the quasi-maximum-likelihood (quasi-ML) detectors of the reverse link system that uses antenna arrays in asynchronous channels when the channel vector is time-invariant and time-varying. A channel vector estimation method based on eigendecomposition for time-invariant channels and its adaptive version suitable for the time-varying channels are also proposed. It is shown that the proposed quasi-ML detector can be regarded as a beamformer followed by a decorrelating filter and that the proposed system performs better than the conventional decorrelator scheme. It is also observed that the performance gain of the proposed scheme over the conventional decorrelator system increases as the numbers of active users and antenna arrays increase.

An efficient implementation of the D-homomorphism for generating de Bruijn (1946) sequences is presented. The number of exclusive-or operations required to generate the next bit for de Bruijn sequences of order n from a de Bruijn function of order k is shown to be approximately k(2/sup W(n-k)/-1), where W(r) is the number of one's in the binary representation of r: therefore, the number of required operations can be reduced to k if the de Bruijn function is selected appropriately. As an application, the stream cipher is indicated.

A vector channel model is proposed and some statistical properties of the asymptotic efficiency of the decorrelating detector with base-station antenna arrays are investigated. It is shown that we can get gain over the conventional decorrelator by employing antenna arrays and the gain increases as the number of antennas increases and the angle dispersion decreases. It is also shown that we can increase the asymptotic efficiency of the decorrelating detector with base-station antenna arrays up to unity if we use an infinite number of antennas when the channel is angle-nondispersive: we cannot, however, increase the asymptotic efficiency of the decorrelating detector by employing base-station antenna arrays if the dispersion is infinite.

A new sequence generation method and a method for reconstructing information symbols for a modulated orthogonal sequence are suggested: the sequence is generated by only integer sums and modular techniques. The autocorrelation and cross-correlation characteristics of the sequence are investigated via a new procedure. A modified sequence also having the orthogonality and satisfying the mathematical lower bound of the cross-correlation is proposed, and the symbol error probability of the sequence is investigated.

Detection of the existence of a desired user is considered in this paper. We assume that the signal to noise ratio is high enough to ignore the effects of noise compared with those of the interference by other users. The inter-user interference and user signals are modeled by the Nakagami model. The observation model for this situation is proposed, the locally optimum test statistic is derived under the model, and the asymptotic performance of the test statistic is compared with that of the envelope detector. We show that the locally optimum detector has performance better than the conventional envelope detector.

This paper considers speech enhancement where the speech signal is modeled with hidden filter models (HFMs), when only noisy speech signal are available. The HFM is a parametric approach for representing the speech waveform in the time domain. We apply the nested EM algorithm for jointly estimating the clean signal and the parameters of HFM and noise model. A computationally efficient implementation of the algorithm is developed by the log-likelihood gradient based on the Kalman filter output in the estimation process. The resulting algorithm does not need framing and may be viewed in the time domain context.

The test statistic of the two-sample locally optimum rank detector is derived for weak random signals in additive noise. It is shown that the two-sample locally optimum rank detector for random signals has interesting similarities to the one-sample locally optimum rank detector for random signals. The detector may also be viewed as a generalization of the two-sample locally optimum rank detector for known signals. Some examples and properties of the score functions constituting the locally optimum rank test statistic are investigated.>

In many applications we have to deal with densities which are highly non-Gaussian or which may have Gaussian shape in the middle but have potent deviations in the tails. To fight against this deviation, we consider Bayesian estimation in the case of non-Gaussian state and measurement noises. The robust estimation problem for linear discrete-time systems is considered here. The non-Gaussian noises are modeled as a mixture. We present a robust recursive estimation model that is an approximate minimum variance estimator with a maximum a posteriori (MAP) decision rule for determining the noise sequence distribution.>

In the previous analysis of the dual sign algorithm (DSA), Gaussian signals were assumed. When the desired response contains additive impulsive interference, however, the analysis seems to be inadequate. In this article, a performance analysis of the DSA is considered when the signals are zero-mean stationary Gaussian, and the additive noise of the desired response is zero-mean stationary contaminated-Gaussian (CG). Through computer simulations, our analysis is validated. It is also shown that the DSA is less vulnerable to impulsive interference than the least-mean square (LMS) algorithm.>

In the article an application of the fuzzy testing of hypothesis to the stochastic-signal detection problem is considered when the signal-to-noise ratio approaches zero. We first obtain the general relationship between the test statistic of the locally optimum fuzzy detector and that of the locally optimum detector. Based on this result, the test statistic and structures of the locally optimum fuzzy detector for stochastic signals are obtained. Several aspects of the locally optimum fuzzy nonlinearity for stochastic signals are also described. Finally, performance characteristics of the locally optimum fuzzy detector are briefly discussed.>

There are two major problems in estimating vocal tract characteristics by conventional linear prediction: estimation accuracy being subject to the characteristics of the excitation source, and the output quality and the estimation accuracy deteriorating with additive background noise. The authors solved these problems as follows: first, estimate the parameters of a robust AR model where the driving noise is a mixture of a Gaussian and an outlier process; then, propose an iterative procedure that involves parameter estimation for uncorrupted speech and data cleaning based on the robust Kalman filter; lastly, the above results are used to enhance speech corrupted by white noise. The results are more efficient and less biased for uncorrupted speech, and superior at low SNR for noisy speech.>

Nonparametric detection of a zero-mean random signal in additive noise is considered. The locally optimum detector based on signs and ranks of observations is derived, for good weak-signal detection performance under any specified noise probability density function. This detector is shown to have interesting similarities to the locally optimum detector for random signals. It may also be viewed as a generalization of the locally optimum rank detector for known signals. Examples of the test statistic of the detector are given for some specific noise probability density functions. Asymptotic and finite sample-size performance of the locally optimum rank detector is also considered.>

In the multipulse linear predictive coding (LPC) speech synthesis, an autoregressive filter is excited by a multipulse excitation consisting of impulses of various amplitudes and locations. The authors statistically model the excitation signal as a zero mean Bernoulli-Gaussian process. The pulse locations are independently distributed with a probability distribution, and the pulse amplitudes are expressed as a Gaussian sequence with zero mean and finite variance. An algorithm is described for estimation of pulse amplitudes and locations based on the Bernoulli-Gaussian process when the number of pulses is given. Results from computer simulation are presented.>

A generalized observation model for signal detection problems is proposed. The model allows consideration of several interesting special cases including additive noise, multiplicative noise, and signal-dependent noise, and includes both deterministic and random signals. Locally optimum (LO) detectors are interesting generalizations of those for the additive noise model. Under the proposed observation model the performance of the LO detectors is compared with that of other common detectors.>

Locally optimum detectors for weak random signals are derived for a generalized observation model incorporating signal-dependent and multiplicative noise. It is shown that the locally optimum random-signal detectors in the generalized observation model are interesting generalizations of those that would be obtained in the purely additive noisy signal model. Examples of explicit results for the locally optimum detector test statistics are given for some typical cases. Both asymptotic and finite sample-size performances of the locally optimum detectors are considered and compared with those of other standard detector structures.>

A method of calculating the maximum entropy per run in binary run-length coding is given when run length is limited to a maximum of M. It is also shown that Huang's bound [1] can be obtained from the present result as M tends to infinity.