Recently, neural networks have been proposed for radar clutter modeling because of the inherent nonlinearity of clutter signals. This paper performs an analysis of the practicality of using a radial basis function (RBF) neural network to model sea clutter and to detect small target embedded in sea clutter. An experiment using an instrumental quality radar was carried out on the eastcoast of Canada to create a rich sea clutter and small surface target database. This database contains both staring and scanning data under various environmental conditions. Using data-sets with different characteristics, we investigate the effects of quantization error, measurement noise, generalization of the neural net over ranges and sampling rate on the RBF clutter model. Despite these physical limitations, the RBF model was shown to approach an optimal predictive performance. The RBF predictor was also applied to detect various small targets in this database based on the constant false alarm rate (CFAR) principle. This RBF-CFAR detector was demonstrated to be able to detect small floating targets even in rough sea conditions.

Estimating a one-dimensional (1-D) chaotic signal in noise is an important problem in chaotic communications and information processing. This problem is theoretically equivalent to the estimation of the initial condition of a chaotic signal. A few studies on this initial condition estimation problem have been carried out for certain specific maps such as the tent map and the logistic map. This problem is investigated for the piecewise linear Markov maps as well as maps that are topologically conjugate to piecewise linear Markov maps. By using the one-to-one correspondence between the initial conditions of a chaotic map and its space of itineraries, several algorithms extending the halving method are developed to estimate the initial condition of a 1-D chaotic signal embedded in additive noise. Performance of these estimators is evaluated using Monte Carlo simulations. At high SNR, the variance of these estimators is found to approach the Cramer-Rao bound.

In the companion paper of Zhou, Yip and Leung (see ibid., vol.47, no.10, p.2655-66, 1999), the maximum likelihood (ML) algorithm for tracking the DOAs of multiple moving targets by passive arrays is presented. In this paper, we provide an asymptotic performance analysis of the algorithm. The statistical consistency of the ML estimates is discussed, and their asymptotic covariances are derived. The Cramer-Rao bounds for the ML estimates are investigated, and their relative efficiency conditions are discussed. The asymptotic performance of the ML tracking algorithm is compared with that of the extended Kalman filter (EKF) under the assumption that the target waveforms are known. Finally, numerical simulation results are used to verify the theoretical results.

This paper describes the combined use of computational intelligence procedures and materials data for monitoring and controlling the growth of thin films using molecular beam epitaxy (MBE). Given ellipsometry data (Ψ and Δ) at a specific wavelength, a genetic algorithm-like method is used to solve an inverse problem, and estimate values of the complex refractive index and the deposition rate. Using a set of such values at different wavelengths, and combining the use of multiwavelength spectroscopic materials data and computational intelligence procedures, it is then possible to provide an optimal estimate of the composition of the material being deposited. Control of the film growth is then accomplished through adjustments of cell temperatures. This procedure is described in this paper, and examples of monitoring and control results are reported for the system of AlxGa1−xAs film on GaAs substrate.

Data fusion is a process dealing with the association, correlation, and combination of data and information from multiple sources to achieve refined position and identity estimates. We consider the registration problem, which is a prerequisite process of a data fusion system to accurately estimate and correct systematic errors. An exact maximum likelihood (EML) algorithm for registration is presented. The algorithm is implemented using a recursive two-step optimization that involves a modified Gauss-Newton procedure to ensure fast convergence. Statistical performance of the algorithm is also investigated, including its consistency and efficiency discussions. In particular, the explicit formulas for both the asymptotic covariance and the Cramer-Rao bound (CRB) are derived. Finally, simulated and real-life multiple radar data are used to evaluate the performance of the proposed algorithm

In this paper, we present an entropy based approach for DOA estimation in Gaussian and non-Gaussian environments. The DOA estimates are obtained by minimizing an entropy measure of the array data in the noise subspace. We show that the entropy approach leads to the MAP algorithm under the Gaussian assumption. Under the non-Gaussian assumption, we apply the varimax norm as an information measure. An intuitive consistency analysis is also performed. Computer simulations are used to demonstrate the effectiveness of the proposed approach

A novel prediction scheme for self-affine fractal signals is presented. The signal is modeled by self-affine linear mappings, whose contraction factors are assumed to follow an auto-regressive (AR) process. In this way, the highly nonlinear time evolution of the fractal signal is captured by the linear AR process of the contraction factors, thereby exploiting the simplicity and ease of computation inherent in the AR model, An adaptive version of the proposed scheme is applied in simulations using the Weierstrass-Mandelbrot cosine fractal, as well as, in practice, using real radar sea clutter data.

A self-calibration DOA estimation algorithm for cyclostationary source signals is presented in which the effects of the sensor gain and phase shift uncertainty have been eliminated. The uniqueness conditions and the asymptotic consistency of the estimates are discussed. An alternating projecting optimization algorithm is provided which lessens the computational load involved in the nonlinear multivariate optimization problem. A numerical example is presented to show the effectiveness of the algorithm.

A high-resolution DOA-estimation technique is proposed to deal with unknown noise-spatial-covariance structure and unknown array-sensor gain. By modelling the source signals as autoregressive moving-average (ARMA) processes with unknown parameters, a formula is derived which relates the source DOAs with the source poles and array-covariance functions. A virtual data matrix is formed, independent of the sensor-gain uncertainty and noise covariance, and a factorisation of this virtual data matrix shows that the subspace-based techniques can be directly applied to estimate the source DOAs. This technique has the advantage that it requires neither the prior knowledge about the sensor-noise covariance nor the sensor-gain calibration. Simulation results are presented to show the effectiveness of the technique and comparisons with the MUSIC algorithm are also included

In this paper, we introduce a preprocessing method for data fusion, based on multiresolution analysis using fractal functions. The reason for choosing this method is that many natural signals belong to the 1/f family and an important class of fractal signals is also of the 1/f type. Because of the self-affinity and the dilation properties, a finite set of fractal interpolation functions (FIF) is chosen for the multiresolution analysis. It is seen that a nested set of subspaces can be generated by the FIF which is equivalent to the set of wavelet subspaces. Through multiresolution analysis, it is possible to reduce the effect of high frequency noise and to keep useful information at the low frequency. Furthermore, such an approach has a localization effect. According to the characteristics of the FIF, the decomposition and reconstruction approach obtained from multiresolution analysis can be implemented by cascade filter banks. Computation complexity is thus also reduced. This method may provide a good way of preprocessing data in fusion.

This paper discusses the problem of registration which is a prerequisite process of a data fusion system to accurately estimate and correct systematic errors. An exact maximum likelihood (EML) registration algorithm is presented. The likelihood criterion is formulated by transforming the measurement data from local sensors to a common system plane. The algorithm is implemented by applying a recursive two-step optimization which involves a modified Gauss-Newton procedure to ensure fast convergence. Numerical simulation studies are conducted to show the effectiveness of the algorithm and comparisons with other registration approaches are provided.

In this paper, we present the results of image compression using smooth block transforms. It is shown that such smooth block transforms solve the blocking effect problem very well, especially for high compression ratio. We also discuss the quasi-optimal smooth block transform, but it leaves an open question--computation complexity.

The authors propose a bandwidth-efficient, asynchronous multiple-access technique with superior probability of error performance. The technique, called waveform division mutiple-access (WDMA), assigns distinguishable baseband waveforms to the transmitters, and transmissions are asynchronous and occupy the entire available bandwidth. In contrast to direct sequence spread spectrum multiple-access (DS/SSMA), each waveform is continuous over the entire signalling interval. The waveforms are selected to achieve a prescribed multiple-access capability and spectral efficiency. The authors postulate a correlation receiver and select the correlating waveform to minimise a worst case average error probability for each receiver, given the transmit waveforms. The requirements of the waveforms are formulated in the context of constrained optimisation. Plots are presented of typical waveforms, spectra, autocorrelation and partial cross-correlation functions. Also presented are error probability curves for noncoherent differential phase shift keying (DPSK) operating in a Gaussian environment. These results indicate that WDMA is a viable alternative to DS/SSMA in applications where bandwidth is scarce and strict security is of no concern. Potential applications are transmission of public safety (police, fire, medical etc.), utility and distress informations.

In this paper, we present a new ATR system for detecting and recognizing targets from a single IR image frame based on neural networks and Gabor functions. It uses Gabor functions to locate potential targets without prior knowledge about their type, size, and orientation. Neural networks are then used to remove false alarms and generate target identification based on information provided by Gabor functions. The new system combines Gabor functions and neural networks in a highly efficient way such that high recognition accuracy rates can be achieved under battlefield conditions. The new system has been successfully tested on hundreds of single frame IR images that contain multiple examples of military vehicles with different size and brightness in various background scenes and orientations, and very high recognition accuracy rates have been achieved.

A high resolution DOA estimation technique in the presence of noise with unknown covariance matrix has been provided. This method provides high resolution at a reasonable computational burden. In this paper, a performance analysis of the DOA estimates obtained by this technique is provided using perturbation theory. Simulation results are presented to compare with the theoretical results. They show very good agreement.

The steered pattern averaging technique (SPAT) for direction-finding in a multiple coherent signal environment is presented. The original array is first partitioned into overlapping subarrays. A transformed array is formulated by applying the weight-and-sum preprocessing scheme on each subarray. The eigenbased algorithms are applied to the transformed array data vector. It is shown that under certain conditions, an effective deconvolution of the coherent signals can be achieved. Computer simulation results are presented to illustrate the effectiveness of the SPAT technique and the improved performance over the spatial smoothing technique

An important problem in high-resolution array processing is the determination of the number of signals arriving at the array. Information theoretic criteria provide a means to achieve this. Two commonly used criteria are the Akaike information criterion (AIC) and minimum descriptive length (MDL) criterion. While the AIC tends to overestimate even at a high signal-to-noise ratio (SNR), the MDL criterion tends to underestimate at low or moderate SNR. By excluding irrelevant parameters, a new log likelihood function has been chosen. Utilizing this new log likelihood function gives a set of more accurate estimates of the eigenvalues and in the establishment of modified information theoretic criteria which moderate the performance of the AIC and the MDL criterion. Computer simulations confirm that the modified criteria have superior performance

The performances of the Akaike (1974) information criterion and the minimum descriptive length criterion methods are examined. The events which lead to erroneous decisions are considered, and, on the basis of these events, the probabilities of error for the two criteria are derived. The probabilities of the first two events are derived based on the asymptotic distribution of the sample eigenvalues of an estimated Hermitian matrix. It is further shown that the probabilities of missing and false alarm for these two criteria can be evaluated to a close approximation. Although the derivation of the probabilities of error is based on an asymptotic analysis, the results are confirmed to be in very close agreement with computer simulation results

Recently, Shenoi and Agrawal [1] discussed the design of low-pass recursive filters using a modified Darlington scheme. A conjecture was made regarding the form of the numerator in the magnitude-squared function. This conjecture was based on two identities satisfied by Chebyshev polynomials. This correspondence provides the proofs for these identities.