Simple adaptive momentum: new algorithm for training multilayer perceptrons

The speed of convergence while training is an important consideration in the use of neural nets. The authors outline a new training algorithm which reduces both the number of iterations and training time required for convergence of multilayer perceptrons, compared to standard back-propagation and conjugate gradient descent algorithms.

Astigmatism in tapered slot antennas

A series of scale model measurements of transverse electromagnetic mode tapered slot antennas are presented. They show that the beam launched by this type of antenna is astigmatic. It is shown how an off-axis spherical mirror can be used to correct this astigmatism to allow efficient coupling to quasi-optical systems. A millimetre wave antenna and mirror combination is described and, with the aid of solid state noise diodes, the coupling of the launched beam to a quasi-optical spectrometer is shown to be in good agreement with that predicted by the scale model measurements.

Adaptive control: an insight

In this article, an overview is given of some of the more common approaches taken in applying adaptive control. Gain scheduling, model reference control and self-tuning control are all discussed and in each case suggestions are given for further reading.

Use of the LMS adaptive signal processing technique to improve signal data rates from practical oil well logging equipment

The simulation and development work that has been undertaken to produce a signal equaliser used to improve the data rates from oil well logging instruments is presented. The instruments are lowered into the drill bore hole suspended by a cable which has poor electrical characteristics. The equaliser described in the paper corrects for the distortions that occur from the cable (dispersion and attenuation) with the result that the instrument can send data at 100 K.bits/second down its own suspension cable of 12 Km in length. The use of simulation techniques and tools were invaluable in generating a model for the distortions and proved to be a useful tool when site testing was not available.

A self-structured adaptive decision feedback equalizer

In a decision feedback equalizer (DFE), the structural parameters, including the decision delay, the feedforward filter (FFF), and feedback filter (FBF) lengths, must be carefully chosen, as they greatly influence the performance. Although the FBF length can be set as the channel memory, there is no closed-form expression for the FFF length and decision delay. In this letter, first we analytically show that the two-dimensional search for the optimum FFF length and decision delay can be simplified to a one-dimensional search and then describe a new adaptive DFE where the optimum structural parameters can be self-adapted.

Equalisation with adaptive time lag

In an adaptive equaliser, the time lag is an important parameter that significantly influences the performance. Only with the optimum time lag that corresponds to the best minimum-mean-square-error (MMSE) performance, can there be best use of the available resources. Many designs, however, choose the time lag either based on preassumption of the channel or simply based on average experience. The relation between the MMSE performance and the time lag is investigated using a new interpretation of the MMSE equaliser, and then a novel adaptive time lag algorithm is proposed based on gradient search. The proposed algorithm can converge to the optimum time lag in the mean and is verified by the numerical simulations provided.

Structure adaptation of linear MMSE adaptive filters

The tap-length, or the number of the taps, is an important structural parameter of the linear MMSE adaptive filter. Although the optimum tap-length that balances performance and complexity varies with scenarios, most current adaptive filters fix the tap-length at some compromise value, making them inefficient to implement especially in time-varying scenarios. A novel gradient search based variable tap-length algorithm is proposed, using the concept of the pseudo-fractional tap-length, and it is shown that the new algorithm can converge to the optimum tap-length in the mean. Results of computer simulations are also provided to verify the analysis.

Adaptive least mean square CDMA detection with Gram–Schmidt pre-processing

The Gram-Schmidt (GS) orthogonalisation procedure has been used to improve the convergence speed of least mean square (LMS) adaptive code-division multiple-access (CDMA) detectors. However, this algorithm updates two sets of parameters, namely the GS transform coefficients and the tap weights, simultaneously. Because of the additional adaptation noise introduced by the former, it is impossible to achieve the same performance as the ideal orthogonalised LMS filter, unlike the result implied in an earlier paper. The authors provide a lower bound on the minimum achievable mean squared error (MSE) as a function of the forgetting factor λ used in finding the GS transform coefficients, and propose a variable-λ algorithm to balance the conflicting requirements of good tracking and low misadjustment.

Constrained surplus energy adaptive blind CDMA detection

The blind minimum output energy (MOE) adaptive detector for code division multiple access (CDMA) signals requires exact knowledge of the received spreading code of the desired user. This requirement can be relaxed by constraining the so-called surplus energy of the adaptive tap-weight vector, but the ideal constraint value is not easily obtained in practice. An algorithm is proposed to adaptively track this value and hence to approach the best possible performance for this class of CDMA detector.

Convergence analysis of chip- and fractionally spaced LMS adaptive multiuser CDMA detectors

This paper analyzes the convergence behavior of the least mean square (LMS) filter when used in an adaptive code division multiple access (CDMA) detector consisting of a tapped delay line with adjustable tap weights. The sampling rate may be equal to or higher than the chip rate, and these correspond to chip-spaced (CS) and fractionally spaced (FS) detection, respectively. It is shown that CS and FS detectors with the same time-span exhibit identical convergence behavior if the baseband received signal is strictly bandlimited to half the chip rate. Even in the practical case when this condition is not met, deviations from this observation are imperceptible unless the initial tap-weight vector gives an extremely large mean squared error (MSE). This phenomenon is carefully explained with reference to the eigenvalues of the correlation matrix when the input signal is not perfectly bandlimited. The inadequacy of the eigenvalue spread of the tap-input correlation matrix as an indicator of the transient behavior and the influence of the initial tap weight vector on convergence speed are highlighted. Specifically, a initialization within the signal subspace or to the origin leads to very much faster convergence compared with initialization in the a noise subspace.

An adaptive linear equalizer with optimum filter length and decision delay

The power of an adaptive equalizer is maximized when the structural parameters including the tap-length and decision delay can be optimally chosen. Although the method for adjusting either the tap-length or decision delay has been proposed, adjusting both simultaneously becomes much more involved as they interact with each other. In this paper, this problem is solved by putting a linear prewhitener before the equalizer, with which the equivalent channel becomes maximum-phase. This implies that the optimum decision delay can be simply ¯xed at the tap-length minus one, while the tap-length can then be chosen using a similar approach as that proposed in our previous work.

Automatic tuning of PID controllers using model reference adaptive control techniques

This paper discusses the application of model reference adaptive control concepts to the automatic tuning of PID controllers. The effectiveness of the proposed method is shown through simulated applications. The gradient approach and simulated examples are provided.