Search

A nonblocking photonic switch can be used to implement a tapped delay line with a large number of adaptive weights and a wide range of time delays. An advantage of using optical tapped delay lines for adaptive filtering is that the operating frequency can be quite high, in the 10-100 GHz range. We present a sparse reconfigurable adaptive filter (SRAF) based on a photonic switch with an input/output connection architecture that can be represented by a matrix of adaptive weights. This unique parallel structure can be reconfigured in an adaptive manner to implement a sparse filter impulse response for use in many applications. We consider an adaptive algorithm for this filter that chooses the input and output delays using a cross-correlation-based approach and connects these delays by weights that are adapted using a gradient algorithm. An alternative adaptive algorithm is also considered that is based on a system identification formulation where the weights are first adapted, and then the appropriate delay combinations are chosen. A search algorithm for implementing the connection constraint required by the SRAF is also discussed whereby each input is connected to only one output at any moment. Computer simulation examples are presented to illustrate the behavior of the filter for a system identification model. Index Terms--Adaptive filtering, photonic switch, sparse filter impulse response, system identification.

The Global Positioning System (GPS) is a one-way satellite-based navigation system employing spread-spectrum techniques that is widely used for commercial and military applications. Although the very low signal-to-noise ratio (SNR) is handled by the large spreading gain, GPS is susceptible to high-power interference signals and various types of jammers. We propose multicomponent receiver architectures for GPS interference suppression. A conventional antenna system is first considered which utilizes a minimum-variance distortionless-response (MVDR) beamformer and assumes that the GPS signal angle of arrival (AOA) and the antenna model are known at the receiver. However, this receiver is sensitive to AOA estimation errors and can have a high computational complexity. This sensitivity problem is eliminated by a multicomponent system based on a multistage matched filter (MF). Since this MF receiver also has a high computational complexity because the jammer AOAs must be estimated, we introduce a blind interference canceler based on the constant modulus (CM) array that is insensitive to AOA estimation errors and has a low computational complexity. Computer simulations are provided to illustrate the performance of the various systems for interference suppression in example signal scenarios.

The Global Positioning System (GPS) was designed to provide location estimates for various civilian and military applications using at least four satellites. Since GPS signals have a low signal-to-noise ratio (SNR), they also have a low signal-to-jammer ratio so that the accuracy of location estimates is influenced by cochannel interference and intentional jammers. We propose a low-complexity blind adaptive receiver that is based on a novel modified despreader and the constant modulus (CM) array. This system is capable of nulling directional interference and capturing the GPS signal of interest without requiring explicit angle-of-arrival (AOA) information. We also consider the multiple satellite problem and extend the proposed receiver to capture several GPS signals of interest. Representative computer simulation examples are presented to illustrate the performance of the multicomponent system for the suppression of different jammer types.

The performance of the linear successive interference canceler (LSIC) is investigated after random spreading sequences are employed. The conditional mean and the signal-to-interference-plus-noise ratio (SINR) of the decision variable in each stage are derived to quantify the effects due to imperfect symbol and amplitude estimates.

A generalized version of the linear successive interface canceller (GLSIC) is described and discussed.

In this paper, we present a two-stage successive cancellation (SC) algorithm that sequentially separates fetal and maternal heartbeats from an intrauterine electrocardiogram (IuECG) signal containing both fetal and maternal QRS complexes. The ECG signal is modeled as a series of fetal, maternal, and noise events. Peak detection is first employed to locate the potential fetal and maternal QRS complexes, referred to as candidate events. Each stage automatically generates a template of a source from the candidate events in the initialization period, and thereafter performs classification of the remaining candidate events based on a template matching technique. The detected events of the stronger signal are subtracted from the composite ECG signal prior to initialization and classification of the weaker signal. Once the fetal and maternal complexes are successfully detected and separated, a counting mechanism is utilized to derive the corresponding heart rates. Computer simulation results on real IuECG data demonstrate the effectiveness of the SC algorithm. Index Terms--Biomedical signal processing, fetal heart-rate estimation, interference cancellation, intrauterine ECG, template matching.

The performance of an adaptive successive interference canceler (ASIC) is examined, using a Wiener model of convergence for the least-mean-square (LMS) algorithm, to determine expressions for the asymptotic mean and variance of the amplitude estimate and the steady-state bit error rate (BER). It is shown that the ASIC's performance exceeds that of the conventional SIC (CSIC), which uses matched filter (MF) output as a received amplitude estimate.

A sequential algorithm that separates cochannel time-division multiple-access signals encountering multipath interference and noise is proposed. Operating on the baseband data obtained from an array of antenna elements, its receiver uses a multistage architecture with each stage comprising of a beamformer and an equalizer. The algorithm overcomes the problem of the beamformer/equalizer's inappropriateness for all the data in a time slot by processing data forward and backward in a sequential noncausal manner.

The blind adaptive constant modulus (CM) array beamformer can separate cochannel signals. The steady-state analyses of a multistage CM Array is examined using a least-mean-square approximation. The adaptive canceler was found to contribute more the overall misadjustments than the adaptive CM beamformer. The transient properties of the system are illustrated via computer simulations.

Cochannel interference occurs when two or more signals overlap in frequency and are present concurrently. Unlike in spread-spectrum multiple-access systems where the different users necessarily share the same channel, cochannel interference is a severe hindrance to frequency- and time-division multiple-access communications, and is typically minimized by interference rejection/suppression techniques. In this paper, rather than using interference suppression, we are interested in the joint estimation of the information-bearing narrow-band cochannel signals. Novel joint estimators are proposed that employ a single-input demodulator with oversampling to compensate for timing uncertainties. Assuming finite impulse-response channel characteristics, maximum likelihood (ML) and maximum a posteriori (MAP) criteria are used to derive cochannel detectors of varying complexities and degrees of performance. In particular, a (suboptimal) two-stage joint MAP symbol detector (JMAPSD) is introduced that has a lower complexity than the single-stage estimators while accruing only a marginal loss in error-rate performance at high signal-to-interference ratios. Assuming only reliable estimates of the primary and secondary signal powers, a blind adaptive JMAPSD algorithm for a priori unknown channels is also derived. The performance of these nonlinear joint estimation algorithms is studied through example computer simulations for two cochannel sources. Index Terms - Blind adaptation, cochannel signal separation, MAPSD, MLSE.

A constant modulus (CM) array algorithm aimed at fast fading environments is proposed. Each source is assumed to have a small angular spread, with the antenna elements closely spaced. The algorithm ensures that a a stage tracks the same source during fast fading conditions, thus producing reliable source estimation. The algorithm is accompanied by a cost function that produces a time-varying step size that slows down the adaptation process during a fade while maintaining the CM property of the original CM array algorithm.

The steady-state observations of the multistage constant modulus (CM) array in the presence of correlated source signals reveal that the signal cross correlation results in phantom sources. Increase of signal leakage in the adaptive cancellers between stages, increases the minimum mean-square error of the later stages. However, a combined cascade/parallel implementation for highly correlated signals can overcome this leakage, which is confirmed by computer simulations. This system operates without any prior information of array configuration but requires an enhanced initialization procedure.

Dual-mode adaptive algorithms with rapid convergence properties are presented for the equalization of frequency selective fading channels and the recovery of time-division multiple access (TDMA) mobile radio signals. The dual-mode structure consists of an auxiliary adaptive filter that estimates the channel during the training cycle. The converged filter weights are used to initialize a parallel bank of filters that are adapted blindly during the data cycle. When the symbol timing is known, this filter bank generates error residuals that are used to perform approximate maximum a posteriori symbol detection (MAPSD) and provide reliable decisions of the transmitted signal. For channels with timing jitter, joint estimation of the channel parameters and the symbol timing using an extended Kalman filter algorithm is proposed. Various methods are described to reduce the computational complexity of the MAP detector, usually at the cost of some performance degradation. Also, a blind MAPSD algorithm for combining signals from spatially diverse receivers is derived. This diversity MAPSD (DMAPSD) algorithm, which can be easily modified for the dual-mode TDMA application, maintains a global set of MAP metrics even while blindly tracking the individual spatial channels using local error estimates. The performance of these single-channel and diversity MAPSD dual-mode algorithms are studied via computer simulations for various channel models, including a mobile radio channel simulator for the IS-54 digital cellular TDMA standard.

A nonlinear system identification model of the required response signal analyzes the behavior of the single-layer backpropagation algorithm for noisy Gaussian training data. Noisy training data has a negligible effect on the ability of the perceptron to learn the underlying weight vector of the training signal model. The weights converge to finite values and the algorithm learns the correct hyperplane related to the system identification method.

A steady-state analysis of the multistage constant modulus (CM) array based on a Wiener formulation of the convergence condition shows that the multistage CM array can recover many cochannel sources and provide estimates of their angles of arrival. The closed-form expressions for the CM array and canceller weight vectors suggest that the unique canceller weights are proportional to the direction vectors of the captured sources. Computer simulations verify the results and demonstrate the impermanence of a cascade adaptive beamforming system for fixed and fading channels.

The use of similar stages of the constant modulus (CM) array and signal canceller allows the recovery of different cochannel sources in a multistage system. When the sources and noise are mutually uncorrelated, the signal canceller eliminates the captured source of the array output. For constant modulus signals, the capture behavior of the CM array is related to the initial weight vector and the relative signal powers at the array output. A Wiener model analyzes the steady-state characteristics of the CM array and the signal canceller.

Stochastic analysis of Rosenblatt's single layer perceptron algorithm for fast learning shows convergence features. Convergence points of the algorithm depend on variance of input signal, and the step size mu, with the algorithm being stable for all mu > 0. The convergence is also studied to find out the effects of noisy perceptron input vectors.

Stochastic convergence studies of second-order adaptive IIR notch filter yield expressions of stability bounds for steady-state coefficient variance and algorithm step size. The analysis depends on extent of gradient linearity near the analytical method for various noise and signal conditions.

In this paper, we describe a high-resolution direction-of-arrival (DOA) estimation algorithm for the enhancement of a lens-based array system. In such a system, the antenna array is followed by an electromagnetic lens and a bank of power detectors (crystal video receivers). This system can handle sources over a large bandwidth and is capable of detecting pulsed sources that exist for short periods of time. The proposed algorithm operates directly on the power estimates provided by a bank of crystal video receivers (CVRs), yielding accurate DOA estimates for multiple narrowband sources. This approach is most effective when the sources are uncorrelated, and it potentially is capable of operating in environments where the signals are highly correlated. We present computer simulations to demonstrate the improved performance achieved by the algorithm for different source scenarios.

A new blind equalization algorithm based on a suboptimum Bayesian symbol-by-symbol detector is presented. It is first shown that the maximum a posteriori (MAP) sequence probabilities can be approximated using the innovations likelihoods generated by a parallel bank of Kalman filters. These filters generate a set of channel estimates conditioned on the possible symbol subsequences contributing to the intersymbol interference. The conditional estimates and MAP symbol metrics are then combined using a sub-optimum Bayesian formula. Two methods are considered to reduce the computational complexity of the algorithm. First, the technique of reduced-state sequence estimation is adopted to reduce the number of symbol subsequences considered in the channel estimation process and hence the number of parallel filters required. Second, it is shown that the Kalman filters can be replaced by simpler least-mean-square (LMS) adaptive filters. A computational complexity analysis of the LMS Bayesian equalizer demonstrates that its implementation in parallel programmable digital signal processing devices is feasible at 16 kbps. The performance of the resulting algorithms is evaluated through bit-error-rate simulations, which are compared to the performance bounds of the maximum-likelihood sequence estimator. It is shown that the Kalman filter and LMS-based algorithms achieve blind start-up and rapid convergence (typically within 200 iterations) for both BPSK and QPSK modulation formats.

Analysis of Rosenblatt's algorithm for single-layer perceptron weight adjustments for two linearly nonseparable data models reveals that the algorithmic weight vector's stationary points are a linear combination of the two input layer-defining vectors. The first data model studied has its basis in a two-layer structure similar to a perceptron. The model's input is an independently structured Gaussian random vector with a zero mean. A single layer structure with a multiplicative output mode yields the second model.

The constant modulus algorithm is widely used for blind equalization in digital communication systems. Its performance function, which indicates the shift of the equalizer output from a prespecified constant, is stochastically evaluated starting with a conditioning step for the transmitted signal. Closed-form equations of the performance function are obtained from a Gaussian process model of the equalizer output and these are analyzed with respect to the probability distribution of the transmitted signals.

The mean-square error (MSE) and probability of correct classification of the single-layer backpropagation algorithm used in the analysis of Gaussian signals are evaluated. Nonlinear functionals of multivariate Gaussian variates are used to evaluate the MSE in terms of a function of the weight vector. Computer simulations of the MSE and the probability of correct simulations showexcellent agreement with existing theoretical data.

The steady-state and transient mean behavior of the single-layer backpropagation algorithm used in adjusting the connection weights of a multilayer perceptron is studied. A Gaussian data vector passing through a linear system defined by the vector of weights is analyzed using a system identification model. Results reveal that a mean weight vector similar to a scaled version of the vector of weights is produced when the algorithm is initialized to the zero vector.