Robert W. Stewart

Design of near perfect reconstruction oversampled filter banks for subband adaptive filters

In this brief, a design algorithm for real-valued and complex-valued oversampled filter banks which yield a low level of inband alias and enable simple subband adaptive structures is presented. The filter banks are either based on complex modulation of a real-valued low-pass prototype or on the direct or modulated setups of real-valued filter banks. If real-valued filter banks are required, then the different channels will have different subsampling ratios so that the bandpass sampling theorem is not violated. This brief also presents design examples of real-valued and complex-valued filter banks.

THE USE OF ACTIVE NOISE CONTROL (ANC) TO REDUCE ACOUSTIC NOISE GENERATED DURING MRI SCANNING: SOME INITIAL RESULTS

MRI scanning generates high levels of acoustic noise that cannot only pose a safety hazard, but also impair communication between staff and patient. In this article we present active noise control (ANC) techniques that introduce antiphase noise to destructively interfere with the MRI noise and with the aim of producing a zone of quiet around the patients ears. Using noise recorded from a 1.0 Tesla midfield MR scanner the acoustic noise generated by three standard MR imaging sequences was replayed to a real time two channel ANC system. The results obtained show a useful attenuation of low-frequency periodic acoustic noise components. Therefore, in combination with standard passive ear protection, this suggests that MR generated acoustic noise can be effectively attenuated at both low and high frequencies leading to improved patient comfort.

Steady-state performance limitations of subband adaptive filters

Nonperfect filterbanks used for subband adaptive filtering (SAF) are known to impose limitations on the steady-state performance of such systems. In this paper, we quantify the minimum mean-square error (MMSE) and the accuracy with which the overall SAF system can model an unknown system that it is set to identify. First, in case of MMSE limits, the error is evaluated based on a power spectral density description of aliased signal components, which is accessible via a source model for the subband signals that we derive. Approximations of the MMSE can be embedded in a signal-to-alias ratio (SAR), which is a factor by which the error power can be reduced by adaptive filtering. With simplifications, SAR only depends on the filterbanks. Second, in case of modeling, we link the accuracy of the SAF system to the filterbank mismatch in perfect reconstruction. When using modulated filterbanks, both error limits-MMSE and inaccuracy-can be linked to the prototype. We explicitly derive this for generalized DFT modulated filterbanks and demonstrate the validity of the analytical error limits and their approximations for a number of examples, whereby the analytically predicted limits of error quantities compare favorably with simulations.

Adaptive IIR filtered-v algorithms for active noise control

This paper reports on the development of “full-gradient” and “simplified-gradient” versions of the filtered-u algorithm for active noise control. After discussing the general principles of active noise control, the paper presents full mathematical derivations of the full-gradient, simplified-gradient, and Feintuch-based versions of the filtered-u algorithm, clearly showing the various levels of simplifying assumptions that are made along the way. Finally, some illustrative simulation results are presented.

Spatial Fading Correlation model using mixtures of Von Mises Fisher distributions

In this paper new expressions for the Spatial Fading Correlation (SFC) functions of Antenna Arrays (AA) in a 3-dimensional (3D) multipath channel are derived. In particular the Uniform Circular Array (UCA) antenna topology is considered. The derivation of the novel SFC function uses a Probability Density Function (PDF) originating from the field of directional statistics, the Von Mises Fisher (VMF) PDF. In particular the novel SFC function is based on the concept of mixture modeling and hence uses a mixture of VMF distributions. Since the SFC function is dependent on the Angle of Arrival (AoA) as well as the power of each cluster, the more appropriate power azimuth colatitude spectrum term has been used. The choice of distribution is validated with the use of Multiple Input Multiple Output (MIMO) experimental data that was obtained in an outdoor drive test campaign in Germany. A mixture can be composed of any number of clusters and this is mainly dependent on the clutter type encountered in the propagation environment. The parameters of the individual clusters within the mixture are derived and an estimation of those parameters is achieved using the soft-Expectation Maximization (EM) algorithm. The results indicate that the proposed model fits well with the MIMO data.

An efficient scheme for broadband adaptive beamforming

This paper introduces an oversampled subband approach to linearly constrained minimum variance adaptive broadband beamforming. This method is motivated by the considerable reduction in computation over fullband implementation and resulting large computational complexity when fullband beamformers with high spatial and spectral resolution are required. We present the proposed subband adaptive beamformer structure, discuss the advantages and limitations of it, and comment on the correct projection of the constraints in the subband domain. In a simulation, the proposed subband structure is compared to a fullband adaptive beamformer, highlighting the benefit of our method.

An automatic sequential recognition method for cortical auditory evoked potentials

The detection of cortical auditory evoked potentials (CAEP), which are part of the electroencephalogram (EEG) in reaction to acoustic stimuli, has important applications such as determining objective audiograms. The detection is usually performed by a human operator, with support from often basic signal processing methods. This paper presents a novel mechanism for the detection of CAEPs, which is fully automatic and stops the measurement when a given confidence is reached. This proposed detector comprises of three stages. First, a feature extraction by a wavelet transform parameterizes the time domain EEG signal by only few transform coefficients. This feature vector is then classified by a neural network which yields a binary vote on every EEG segment. Finally, a sequential statistical test is performed on successive classifications; this stops the measurement if a specified decision confidence has been reached. The adjustment of the detector according to a clinical database is discussed. Thus adjusted, the proposed CAEP detection scheme is applied to a study, and compared with a human operator. The results demonstrate that this method can attain similar results, but outperforms the human expert for stimulation levels close to the hearing threshold.

Theory and applications of adaptive second order IIR Volterra filters

An adaptive nonlinear filter based on a second order Volterra series and on an IIR filter structure is presented. This filter is able to model higher than second order nonlinearities for systems where the nonlinearities are harmonically related. This solution represents an alternative to using higher than second order Volterra filters. We present a full derivation of this gradient search based adaptive nonlinear filter and also highlight the various assumptions and simplifications which require to be made in order to produce a practical algorithm. A comparison is made in terms of the performance and computational complexity between an adaptive second order IIR Volterra filter and an adaptive second and third order Volterra filters.

High-order system identification with an adaptive recursive second-order polynomial filter

In this letter, an adaptive recursive nonlinear filter based on the Volterra series and an infinite impulse response (IIR) structure is considered. For certain types of nonlinear systems where high-order nonlinearities are recursively generated, we show that the adaptive recursive second-order polynomial filter has improved performance over the well-known (nonrecursive) adaptive second-order Volterra filter and a third-order Volterra filter. This filter represents an alternative to using a traditional Volterra filter whose order has been increased to match that of the system being modeled.

UHF white space network for rural smart grid communications

We present a white space communications test bed running in the Scottish Highlands and Islands, and discuss its feasibility for smart grid communications. The network aims to serve communities that have great potential for distributed generation of electricity, by means of wind, water, and tidal power. However, smart grid applications such as remote meter reading and load balancing are impaired by the scarcety or lack of communications infrastructure in remote rural areas such as the Scottish Highlands and Islands. We argue that the proposed system is based on a network of energy self-sufficient radio relay nodes that make it a robust and independent medium to support smart grid communications in rural settings.