We consider a wireless system where multiple users are served in Downlink (DL) by one Base Station (BS) and one Relay Station (RS). Previous research has shown that, combining two users' messages via Superposition Coding (SC) could enhance their achievable rate and fairness, in both two-user and multi-user systems. In this work, we propose two novel SC schemes referred as Relayed/Direct (RD-SC) and Relayed/Relayed SC (RR-SC) schemes, that combine the messages to a pair of Relayed/Direct users or Relayed/Relayed users into three or four SC layers, where the number of SC layers is equal to the number of available links. For each scheme, power allocation under sum-rate maximization is analyzed, enabling to derive the optimal power ratios to each SC layer numerically. Then, we design the non-orthogonal Max Rate-Two User-SC (MR-TU-SC) Scheduler and Proportional Fair-Two User-SC (PF-TU-SC) Scheduler where a pair of selected users are allocated simultaneously based on RD-SC and RR-SC schemes. The simulation results show that the proposed schedulers significantly outperform the conventional orthogonal schedulers where a single user is allocated on a unit resource block, in terms of system throughput, fairness and outage, while approaching the upper bound performance.

We consider the problem of femtocell/macrocell interference mitigation in an OFDMA based Downlink (DL) heterogeneous system. Most conventional methods rely on the information of the interference channel states and Macrocell User (MU) allocation provided by a control channel, causing large overhead increase. In our proposed method, each Femtocell Base Station (FBS) is able to predict the subchannel allocation of MUs based on partial Channel State Information (CSI) feedback and overhearing from the MUs to the Macrocell Base Station (MBS). Based on this estimation, each FBS performs subchannel and power allocation to Femtocell Users (FUs). With this procedure, the proposed method achieves excellent MU/FU sum-rate while greatly reducing required overhead compared to reference schemes.

Anonymous tag set cardinality estimation problem of Radio Frequency IDentification (RFID) using Maximum A Posteriori (MAP) approach is studied in this paper. The posterior probability of the total number of tags, given the frame size and the observed number of non-empty slots, is firstly determined. Then, the total number of tags is estimated to maximize the posterior probability. Computer simulation results demonstrate the effectiveness of the proposed approach.

We consider the problem of Downlink (DL) power allocation at a Femtocell Base Station (FBS) in a macro-cell/femtocell coexisting system based on OFDMA transmission. To mitigate the interference to macrocell users, many conventional methods rely on the information of the interference channel states and macrocell user allocation provided by a control channel, causing large overhead increase. Instead, in our proposed method, each FBS is able to predict the subchannel allocation of macrocell users based on partially overheard Channel State Information (CSI) from the macrocell users to the Macrocell Base Station (MBS). Based on this estimation, the FBS performs subchannel and power allocation to femtocell users. With this procedure, the proposed method can mitigate the interference to macrocell users without largely increasing the control overhead. Simulations show the validity of the proposed method, providing sum-rate gains for both femtocells and macrocell users compared to reference schemes.

The paper proposes a subcarrier power allocation method for downlink OFDMA systems. The proposed method utilizes an iterative water-filling (IWF) algorithm with signal-to-leakage-plus-noise ratio (SLNR) rather than conventional signal-to-interference-plus-noise ratio (SINR), which enables us to considerably reduce the overhead. Computer simulation results show that the proposed IWF based on SLNR with 1-shot 1-bit feedback per subcarrier can achieve very close sum-rate performance to the conventional IWF based on SINR, while only replacing SINR to SLNR in the IWF algorithm results in poor performance.

A relay system where two users are served by one Base Station (BS) and one Relay Station (RS) is focused on. A scheme based on Superposition Coding (SC) for Decode-and-Forward (DF) half-duplex relaying is proposed. Thanks to its design and power allocation ratios derived analytically, the proposed scheme is able to serve both users simultaneously, while attaining the same sum rate as the optimal benchmark scheme that exclusively allocates the best user. Simulation results show considerable fairness improvement over a large range of SNRs.

A Superposition Coding (SC) scheme using discrete Adaptive Modulation (AM) for a three-node wireless relay system is proposed. In the proposed method, 2/4-QAM and 4/16-QAM hierarchical modulations are used to generate the SC message composed of a basic and superposed messages. We derive the analytical throughput expression of the system and determine the necessary conditions for guaranteeing optimal power allocation. The simulation results show that the proposed power allocation achieves a throughput close to the one obtained by exhaustive search. Moreover, over a large range of SNRs, the proposed method largely outperforms the conventional relaying methods such as multi-hop and cooperative diversity, proving the substantial benefits of SC even under practical discrete AM and decoding errors during SIC.

Optimized resource allocation of the Downlink (DL) in wireless systems utilizing Multi-Carrier (MC) transmission requires Channel State Information (CSI) feedback for each user/subchannel to the Base Station (BS), consuming a high amount of Uplink (UL) radio resources. To alleviate this problem, several works have considered contention-based CSI feedback in the UL control channel. We propose such a feedback scheme for a generic MC system, based on the idea of variable collision protection, where the probability that a feedback information experiences a collision depends on its importance. By partitioning the CSI into orthogonal layers of priority, and allocating different numbers of feedback slots to each layer, this scheme ensures that the feedback success probability is higher for the CSI with better quality, which is more likely to be used by the scheduler. Furthermore, we present a theoretical performance analysis of the proposed scheme, assuming Maximum CSI (Max CSI) and normalized Proportional Fair Scheduler (PFS), where a tight approximation of the achievable throughput is obtained assuming discrete Adaptive Modulation (AM) and CSI feedback which are relevant for the practical systems. Analytical and simulation results show that our proposed scheme provides an excellent trade-off between system performance and feedback overhead.

We propose an application of variational Bayesian inference to the problem of blind estimation of single-input multiple-output (SIMO), finite-impulse-response (FIR) channels. The blind channel estimation problem has been extensively studied well over a decade. Nevertheless, there are some types of channel that cannot be successfully estimated by existing blind methods. One example of such a channel type is frequency-selective, exponentially-weighted Rayleigh fading channel. Our blind estimation approach to it is to incorporate prior knowledge on channel weighting in the form of channel prior distribution, then obtain the approximation to the posterior distribution of the channel using the variational Bayesian framework. Despite the error due to the mean field approximation involved in the variational framework, the derived algorithm has been verified by simulation to provide much improvements over the existing ones in a statistical sense.

We propose an application of the Expectation-Maximization (EM) algorithm to the problem of blind estimation of single-input multiple-output (SIMO), finite-impulse-response (FIR) channels. We first assume Gaussian input to formulate an EM-based estimation of the signal subspace of the output covariance matrix. This Gaussian assumption allows us to utilize knowledge from EM-based probabilistic principle component analysis (P-PCA). Next, we show that the equilibrium point of the EM iteration equations is reached without the Gaussian assumption, which suggests usage of non-Gaussian communication input signals. The estimated signal subspace is then utilized to identify the channels. In principle, the proposed method yields the same channel estimates as the widely-known subspace method, but is computationally more efficient. In addition, unlike typical EM applications, the proposed scheme is free from cumbersome parameter initialization issue, which greatly increases flexibility of the proposed scheme.

One limiting issue in implementing high-speed wireless systems is the impairment associated with analog processing due to component imperfections. In uplink transmission of multiuser systems, a major source of such impairment is IQ imbalance (IQI) introduced at multiple transmitters. In this paper, we analyze the effect of the transmitter (Tx) IQI on OFDMA receiver. To cope with the inter-user interference problem due to Tx IQI, we propose a novel subcarrier allocation scheme, which has high tolerance to such Tx IQ distortion.

This paper proposes a transmit beamforming and subcarrier power allocation method for orthogonal frequency division multiple access (OFDMA) systems based on ¿signal-to-leakage-plus-noise ratio¿ (SLNR). As the transmit beamforming vector control criterion, we employ the maximization of the SLNR at each base station, while the subcarrier power allocation is performed by iterative water-filling algorithm using the SLNR of each subcarrier. The SLNR based approach enables us to obtain closed form beamforming vector, and achieve power allocation without sending any signal to the mobile terminal. We also discuss the validity of the SLNR based beamforming vector in terms of Pareto optimality. Computer simulations show the promising performance of the proposed method and the validity of the analysis of the SLNR based beamforming vector.

This paper considers a transmit beamforming and subcarrier power allocation method for orthogonal frequency division multiple access (OFDMA) systems. As the beamforming vector control criterion, we employ the maximization of so-called signal-to-leakage-plus-noise ratio (SLNR) at each base station, which enable us to obtain closed form beamforming vector by using only locally available information. We also discuss the local optimality of the SLNR based beamforming vector. As for the subcarrier power allocation, two different approaches are employed, namely, the equalization of signal-to-interference-plus-noise ratio (SINR) for subcarriers and the maximization of sum rate of subcarriers. Computer simulation results show the validity of the transmit beamforming and power allocation method with highlighting the difference between the two power allocation algorithms.

Abstract In this paper, we propose several power allocation schemes for multi‐input multi‐output (MIMO) orthogonal frequency division multiplexing (OFDM) transmission based on the minimization of an approximated bit error rate (BER) expression, and we evaluate the different solutionsviafield trial experimentations. The methods illustrated in this paper, serve to allocate power among the different transmit antennas and the different subcarriers which compose the MIMO OFDM transmitted signal. Several solutions are available to perform power allocation. Frequency domain power allocation, spatial domain power allocation and combined spatial and frequency power allocation are evaluated. We first review and describe the analytical solution for each power allocation scheme and then evaluate the complexity in terms of both computational operations and BER performances. Simulation results show the performance in term of BER and link the advantage of each possibility of power distribution with the associated complexity. Copyright © 2009 John Wiley & Sons, Ltd.

This paper proposes co-channel interference cancellation / equalization methods for downlink block transmission with cyclic prefix in multi-cell environments using multiple receiving antennas and iterative least-square (LS) interference cancellation. Due to circulant structure of channel matrices in block transmission schemes with cyclic prefix, pseudo inverse matrices, which is required in the iterative LS method, can be obtained with low computational complexity and is multiplied to the received signal in an efficient manner. Computer simulation results show that the proposed iterative LS interference cancellation and equalization method can achieve almost the same bit error rate (BER) performance as an minimum mean-square-error (MMSE) approach with lower computational complexity and without noise variance estimation.

We propose generic relay and subcarrier allocation schemes for multicarrier (MC) system with amplify-and-forward (AF) relays. The outage probability bounds are derived analytically for each scheme. Simulation results show that these bounds are very tight and better than the bounds obtained straightforwardly from the analysis in the Single-Carrier (SC) case. This is because in our analysis we reckon with the increased degree of freedom brought by the parallel channels. One of the proposed schemes, the Average Best Relay Selection scheme, is best suited for practical implementation since it approaches the best performance while minimizing the required amount of signaling.

This paper proposes uplink channel estimation schemes for OFDMA system. If frequency responses of assigned subcarriers between corresponding users and base station (BS) are put together into one frequency selective fading channel, time domain expression of the channel, in other words, overall channel impulse response has a significant property, namely, only about the same number of elements as the length of CP have nonzero values. Based on the analysis, we propose a least square channel estimation scheme and a simple remedy for conventional channel estimation scheme. Computer simulation results show that the proposed schemes can significantly improve BER performance.

This paper proposes a new adaptive filter algorithm for system identification using independent component analysis (ICA), which separates the signal from noisy observation under the assumption that the signal and noise are independent. We first introduce an augmented state-space expression of the observed signal, representing the problem in terms of ICA, and then use an adaptive gradient descent algorithm to separate the noise from the signal. A local convergence condition is also shown. The proposed algorithm can be applied to the acoustic echo cancellation problem directly and some simulations have been carried out to illustrate its effectiveness.

In this paper, we propose to adapt simultaneously the transmit power both in the spatial and frequency domains using a heuristic expression of the bit error rate (BER) for each sub- carrier and transmit antenna. The proposed method consists of grouping a certain number of subcarriers and performing local power adaptation in each subcarrier group and transmit antenna. The subcarrier grouping is performed in such a way that equalizes the average channel condition of each subcarrier group. The grouping and the local power adaptation allow us to take advantage of the channel variations and to reduce the computational complexity of the proposed power distribution scheme. With the simplicity of the heuristic BER expression, we can obtain a closed form expression of the transmit power to be allocated. Simulation results show significant performance gain in term of BER compared to the equal power distribution.

In this paper, we develop a power distribution scheme for multi-input multi-output (MIMO) multi-carrier (MC) code division multiple access (CDMA) systems with two-dimensional spreading for wireless transmission. We propose to allocate power among symbols of the active users transmitted from multiple transmit antennas based on a heuristic expression of the bit error rate (BER). Optimization principle consists of minimizing the average BER of all users in term of channel conditions, user multiplexing and number of transmit antennas. Due to the simplicity of the approximate BER expression, we can obtain a closed form expression of the optimum power to be allocated to each user on each transmit antenna. Simulation results show significant improvement of the BER performances compared to the equal power distribution for various modulations, several encoding gains and two antenna configurations.

This paper investigates downlink channel estimation schemes for multi-cell block transmission systems with cyclic prefix and proposes a time domain channel estimation scheme using iterative least square interference cancellation. The proposed scheme takes advantage of the fact that each received pilot signal exists in up to the length of the cyclic prefix pulse one dimensional subspace in the fast Fourier transform (FFT) size dimensional received signal vector space. Computer simulation results show that the proposed scheme can achieve almost the same channel estimation accuracy in multi-cell scenario as that in single-cell environment with small number of iterations.

The objective of this letter is to analyze the effects of frequency mismatch for an adaptive algorithm that becomes the Filtered-X LMS algorithm when the reference signals are purely sinusoidal. The Filtered-X LMS algorithm is often used for active control of acoustic noise. For the case of sinusoidal noise sources, if there is a deviation between the frequency used in the adaptive algorithm and its true value (frequency mismatch), the performance of the Filtered-X LMS algorithm might degrade considerably. In this letter, using the equivalent transfer function method, the effects of frequency mismatch are analyzed precisely. Finally, computer simulations are presented to demonstrate the obtained results

It is well known that the conventional adaptive algorithm for cancellation of the effect of the feedback path in hearing aids is biased. In this paper the stationary point of a modified adaptive algorithm by splitting the delay in the conventional one is derived by a direct method considering the causality constraint and an explicit expression of the bias is obtained. Also, a stability condition is mentioned. Then, when the input signal to hearing aids is a moving average (MA) process or an autoregressive (AR) process, unbiased algorithms for identification of the feedback path are proposed. It is shown by simulations that the new algorithm for an AR process converges considerably faster than that of the previously proposed two-channel adaptive algorithm.

Multi-carrier modulation (MCM) scheme such as orthogonal frequency division multiplexing (OFDM), is a popular method for high data rate wireless transmission. This paper proposes a novel transmit power allocation for MCM. This technique consists of adapting the power allocation in the frequency domain by considering the channel variations. Optimization process is based on the optimality of the global bit error rate (BER). Simulation results show significant gains for several modulations and setting configurations

This paper proposes a simple detection scheme for a single carrier block transmission (SCBT) without guard interval (GI). Although the SCBT without the GI can significantly increase bandwidth efficiency, the received signal suffers from interblock interference (IBI) and intersymbol interference (ISI), which cannot be effectively equalized by a frequency domain equalizer (FDE). In the proposed scheme, the IBI is cancelled by using a previously detected signal. Moreover, taking advantage of the temporal localization of the difference of the ISI component between the received signal with and without CP, we change the channel matrix from Toeplitz structure into a circulant matrix in order to utilize the conventional FDE. Computer simulation results show that the proposed scheme can outperform a linear MMSE equalizer with lower computational complexity

This paper proposes a burst noise cancellation scheme for single carrier block transmission with cyclic prefix (SC-CP) systems. In the proposed scheme, we simply force the received signals which are collapsed by the burst noise to be zeros. Since this processing introduces an inter-symbol interference (ISI), which cannot be effectively equalized with a conventional frequency domain equalizer (FDE), we also propose an ISI cancellation scheme. We firstly generate a replica signal of the ISI using tentative decisions in order to make the defective channel matrix to be circulant, and then the conventional frequency domain equalization is performed to compensate the ISI. Also, we have utilized newly derived minimum mean-square-error (MMSE) based FDE weights for the replica signals generation. Computer simulation results show that the proposed burst noise cancellation scheme with the ISI cancellation can significantly improve the bit error rate (BER) performance even when almost 10% of the received signals are collapsed by the burst noise

In the presence of tonal noise generated by periodic noise source like rotating machines, the filtered-X LMS (FXLMS) algorithm is used for active control of such noises. However, the algorithm is derived under the assumption of slow adaptation limit and the exact analysis of the algorithm is restricted to the case of one real sinusoid in the literature. In this paper, for the general case of arbitrary number of sources, the characteristic polynomial of the equivalent linear system describing the FXLMS algorithm is derived and a method for calculating the stability limit is presented. Also, a related new algorithm free from the above assumption, which is nonlinear with respect to the tap weights, is proposed. Simulation results show that in the early stage of adaptation the new algorithm gives faster decay of errors.

We exhibit an important property called the asymptotic equipartition property (AEP) on empirical sequences in an ergodic multiagent Markov decision process (MDP). Using the AEP which facilitates the analysis of multiagent learning, we give a statistical property of multiagent learning, such as reinforcement learning (RL), near the end of the learning process. We examine the effect of the conditions among the agents on the achievement of a cooperative policy in three different cases: blind, visible, and communicable. Also, we derive a bound on the speed with which the empirical sequence converges to the best sequence in probability, so that the multiagent learning yields the best cooperative result.

Often one encounters the presence of tonal noise in many active noise control applications. Such noise, usually generated by periodic noise sources like rotating machines, is cancelled by synthesizing the so-called antinoise by a set of adaptive filters which are trained to model the noise generation mechanism. Performance of such noise cancellation schemes depends on, among other things, the convergence characteristics of the adaptive algorithm deployed. In this paper, we consider a multireference complex least mean square (LMS) algorithm that can be used to train a set of adaptive filters to counter an arbitrary number of periodic noise sources. A deterministic convergence analysis of the multireference algorithm is carried out and necessary as well as sufficient conditions for convergence are derived by exploiting the properties of the input correlation matrix and a related product matrix. It is also shown that under convergence condition, the energy of each error sequence is independent of the tonal frequencies. An optimal step size for fastest convergence is then worked out by minimizing the error energy.

In this paper, we regard the sequence of returns as outputs from a parametric compound source. Utilizing the fact that the coding rate of the source shows the amount of information about the return, we describe l-learning algorithms based on the predictive coding idea for estimating an expected information gain concerning future information and give a convergence proof of the information gain. Using the information gain, we propose the ratio w of return loss to information gain as a new criterion to be used in probabilistic action-selection strategies. In experimental results, we found that our w-based strategy performs well compared with the conventional Q-based strategy.

An efficient scheme is presented for implementing the LMS-based transversal adaptive filter in block floating point (BFP) format which permits processing of data over a wide dynamic range at a processor cost marginally higher than that of a fixed point processor. Appropriate BFP formats for both the data and the filter coefficients have been adopted and adjustments made in filtering as well as weight updating operations in order to sustain the adopted format and also to prevent overflow in both these operations jointly. For the presented method to work properly, the algorithm step size is to be chosen below an upper limit, which is, however, not very restrictive when compared with the upper bound for convergence, thereby having marginal effect on convergence speed.

The paper presents some convergence results of an adaptive filter algorithm for a feedback-type active noise control (ANC) system with feedback path modeling using a dither signal and for feedback path modeling of hearing aids. In these systems, there is an adaptive filter in the feedback path so that the conventional method for analysis does not work. A frequency domain method is used for the analysis. It is based on the averaging method and the frequency domain expression of the adaptive algorithm with a causality constraint. The convergence conditions are derived for both problems and a bias expression of the weights for the latter problem is presented. The positive realness of the transfer function of the optimal predictor of the incoming signal is required for stability. Finally, the convergence condition and the bias are explicitly given for simple examples and their validities are shown by some simulations.

The hybrid active noise control (ANC) system consists of the combination of feedforward and feedback type ANCs. The filtered-X LMS algorithm is used to adapt the controllers.. The averaging method is applied to the hybrid ANC system with the frequency domain expression to derive a stationary point of the adaptive algorithm. Then, it is found that though the adaptive filters of both ANCs are operating simultaneously, the stationary point of each adaptive filter is same with that for the case where each ANC system is operating separately. Then the local convergence condition is derived. Finally, the convergence condition is explicitly examined for a simple case and its validity is shown by some simulations.

The feedback-type active noise control (ANC) system uses only one microphone to provide necessary signals for adjusting the adaptive filter. Due to the complicated nature of the whole adaptive filter structure there have been no theoretical results about its convergence properties. First a stationary point of the adaptive filter using the filtered-X LMS algorithm is obtained by the averaging method combined with the frequency domain technique. Then the local convergence condition is derived. This is a counterpart of the well-known 90/spl deg/ condition for the feedforward-type ANC. Finally, the convergence condition is explicitly given for a simple example and its validity is shown by some simulations.

The upper limit of the step size and the variance of error signal for a two-band delayless subband adaptive filter are theoretically evaluated in the frequency domain. The averaging method and the ordinary differential equation (ODE) method are used for this purpose. It is shown that the calculated result of upper limit of the step size has a similar form of that of the fullband LMS adaptive filter. In the ODE analysis, not only the convergence condition for the adaptive filter but also the variance of the error signal can be derived. The proposed method is also applied to a half-band delayless subband adaptive filter. We demonstrate that these theoretical results give good approximations for simulation results.

The normalized LMS (N-LMS) algorithm has a disadvantage that the convergence rate is much worse when the input signal is colored. To overcome this, the affine projection algorithm and the block orthogonal projection (BOP) algorithm which applied the block signal processing technique to the N-LMS algorithm are proposed although the reason why they are tough against the colored signal is not given yet. This paper gives the convergence rate of the BOP algorithm for colored input signals, which shows the superiority of the BOP algorithm. To put it concretely, we derive the expression of the convergence rate, propose an approximation method to calculate it, and confirm the result by computer simulations. We also consider the relation between the block size and the convergence rate formally and geometrically.

Delayless subband adaptive digital filters (ADFs) have been proposed to avoid the drawbacks of the conventional subband ADF, which are transmission delay associated with the filterbank and degradation of the performance due to aliasing effects. In the delayless subband ADF, the tap-weight vectors are computed in subbands and then they are transformed into the tap-weight vector in fullband by a certain transform. In this letter, a general formulation for the delayless subband ADF is proposed, and some requirements on the fullband/subband transform are investigated.

An analysis of a delayless subband adaptive digital filter (ADF) structure is presented. In this structure adaptive weights in each subband are computed by the LMS algorithm and then transformed into those in fullband by the Hadamard transform. The conventional subband ADF has transmission delay and aliasing effects associated with the filter bank. However in this manner such defects are avoided while retaining the computational and convergence speed advantages of subband decomposition. In addition the overall transfer function of the novel type of subband ADF is strictly equivalent to that of the Wiener filter for the fullband ADF. Also, a characteristic equation is derived to discuss the stability of the adaptation algorithm. Some numerical results show the good performance of this scheme.

This paper presents an analysis of a subband adaptive filter structure using noncausal Wiener filter theory and multirate system concepts. The transfer function of the overall system is close to that of the full band optimal Wiener filter whereas that of the conventional subband filter contains a delay of the filter bank.

Generally, the output of a filter bank for a stationary input signal is not stationary but cyclostationary. In this paper, by using cyclostationary spectral analysis, the spectral correlation density of this output is derived. Using this result we derive a criterion to construct an optimal 2-band perfect reconstruction filter bank which minimizes the averaged mean squared reconstruction error when the high pass band signal is dropped. By adding constrains to the filter coefficients, the biorthogonal filter bank, the conjugate quadrature filter bank and the biorthogonal linear phase filter bank are respectively obtained. Some numerical results are also presented for optimal biorthogonal and PR linear phase filter banks which are compared in terms of some performance measures.

A new vectorized systolic array for computing the weights of the block recursive least squares (RLS) with p linear constraints is proposed. Using the idea of M-invariant matrices, an upper triangular array for updating R/sub n/ of the modified QR decomposition of the data matrix is first derived where the first p rows corresponding to the constraints are in a frozen mode. Next, a lower triangular array for updating R/sub n//sup -T/ is derived where the first p rows are in no operation mode. Connecting these two arrays, the overall near rhombic array is obtained for computing the weights. Also, the numerical stability of the algorithm is examined by simulations.>

A new vectorized rhombic systolic array for computing tap coefficients in the block recursive least squares (RLS) method is proposed. The array is based on the inverse factorization method of Pan and Plemmons (1989) and on the triangular array for computing residual vectors in block RLS using the Householder transformation of Liu et al. (1990). The new array is obtained by generalizing the inverse factorizations for the block case and then examining the relation between the a priori error vector and the a posteriori error vector to get rid of the matrix operations.>

The authors present a novel nonlinear regulator design method that integrates linear optimal control techniques and nonlinear neural network learning methods. Multilayered neural networks are used to add nonlinear effects to the linear optimal regulator (LOR). The regulator can compensate for nonlinear system uncertainties that are not considered in the LOR design and can tolerate a wider range of uncertainties than the LOR alone. The salient feature of the regulator is that the control performance is much improved by using a priori knowledge of the plant dynamics as the system equation and the corresponding LOR. Computer simulations are performed to show the applicability and the limitations of the regulator.

The authors present a nonlinear regulator design method that uses multilayered neural networks (MNNs) for nonlinear systems operating in the presence of system uncertainties. The regulator consists of two MNNs for modeling and control in addition to the classical linear optimal regulator. The MNN for modeling is connected in parallel with the system equation to compensate for system uncertainties. The MNN for control is connected in parallel with the linear optimal regulator to compensate for the control errors caused by system uncertainties. Both MNNs are trained so that a desired response of the plant is obtained. A training signal of the MNN for control is acquired through the results of modeling. It is shown that the regulator can generate more sophisticated control inputs than the linear optimal regulator, even when the MNNs get stuck at local minima. A priori knowledge of the plant dynamics can be incorporated into the system equation and into the corresponding linear optimal regulator. The learning time is greatly reduced as compared with the case where MNN is used alone for modeling or control. Numerical simulations show that the regulator can stabilize a plant that is not stabilized by the linear regulator

In this paper, first, the normalized Triangular Lattice and Escalator algorithms for recursive least squares estimation are presented. Next, roundoff errors of both algorithms in fixed point arithmetic are analyzed and compared. Expressions are derived for the biases in the estimated coefficients due to roundoff errors. By improving the method of analysis, better expressions are obtained, as compared with the previous ones.

This paper is concerned with the estimation of frequencies of sinusoids based on the data disturbed by colored noise. The noise is assumed to be an MA process and the bias compensated least squares (BCLS) algorithm is derived for estimating the coefficients of the linear difference equation for the sum of sinusoids. For the white noise case, the algorithm is closely related to the well-known Pisarenko method. To see the effectiveness of our method, several simulations are performed and the spectral estimates by ours and Burg method are presented.

After a review of the statistical properties of multivariate autoregressive (MAR) spectral analysis method, we derive the asymptotic distributions of coherence and power contribution ratios obtained by MAR modeling. Then, using the results, we devise a statistical test about the causality relations between variates. Finally we apply the test to simulation and actual data to see the validity of the present analysis.

This paper presents two methods for AR spectrum analysis based on a noisy autocovariance sequence. Unlike most of the traditional spectrum analysis methods, we assume that there is an observed autocovariance sequence, possibly with errors. Such a situation occurs in the Fourier spectroscopy. We fit AR models to this autocovariance sequence by two methods. The one is based on the least squares fit of the autocovariances to the Yule-Walker equations and the other is based on the maximum likelihood estimation of partial autocorrelation coefficients by nonlinear optimization techniques. We apply these two methods to simulated and real plasma data and compare their results.

Optimal linear coding schemes for feedback communication with noisy side information are derived. The performance of two schemes with side information are compared with the information-theoretic bounds. The side information at the transmitter Significantly reduces the mean-squared error. The corresponding continuous-time system is briefly mentioned.

A feedback communication system which has noisy side information at the receiver is considered. The estimation scheme is a simple extension of the conventional one without side information. The resulting mean-squared error is compared with the rate-distortion bound obtained recently by Wyner and Ziv.