Optical precursors and Beer's law violations; non-exponential propagation losses in water

We develop a model for exponential decay of broadband pulses, and examine its implications for experiments on optical precursors. One of the signature features of Brillouin precursors is attenuation with a less rapid decay than that predicted by Beer's Law. Depending on the pulse parameters and the model that is adopted for the dielectric properties of the medium, the limiting z-dependence of the loss has been described as z(-1/2), z(-1/3), exponential, or, in more detailed descriptions, some combination of the above. Experimental results in the search for precursors are examined in light of the different models, and a stringent test for sub-exponential decay is applied to data on propagation of 500 femtosecond pulses through 1-5 meters of water. (C) 2005 Optical Society of America.

Fatigue crack propagation in a sintered 2xxx series aluminium alloy

Many potential applications for sintered aluminium are limited by the poor fatigue properties of the material. In order to increase understanding of the fatigue mechanisms in sintered aluminium, fatigue tests were carried out on a sintered 2xxx series aluminium alloy, AMB-2712. The alloy has a fatigue endurance strength of approximately 145 MPa (R = 0.1). Three regions were identified on the fatigue fracture surfaces. Region I contains the initiation site and transgranular crack propagation. When the size of the cyclic plastic zone ahead of the crack becomes comparable to the grain size, microstructural damage at the crack tip results in a transition to intergranular propagation. Region 2 mainly contains intergranularly fractured material, whilst the final fracture area makes up Region 3, in the form of dimple coalescence and intergranular failure. Transgranular fractographic features observed on fatigued specimens include fissure-type striations, cross-hatched grains, furrowed grains and grains containing step-like features. (c) 2006 Elsevier B.V. All rights reserved.

Effect of line of sight propagation on capacity of an indoor MIMO system

In this paper the performance of a multiple input multiple output (MIMO) wireless communication system operating in an indoor environment, featuring both line of sight (LOS) and non-line of sight (NLOS) signal propagation, is assessed. In the model the scattering objects are assumed to be uniformly distributed in an area surrounding the transmitting and receiving array antennas. Mutual coupling effects in the arrays are treated in an exact manner. However interactions with scattering objects are taken into account via a single bounce approach. Computer simulations are carried out for the system capacity for varying inter-element spacing in the receiving array for assumed values of LOS/NLOS power fraction and signal to noise ratio (SNR).

Demonstrating the effects of multi-path propagation and advantages of diversity antenna techniques

In this paper, experimental investigations are performed into assessing the quality of communication link between Bluetooth devices in an indoor environment, as an initial step of demonstrating benefits of diversity and smart antenna techniques in mobile computing.

Adaptive back-propagation in on-line learning of multilayer networks

An adaptive back-propagation algorithm is studied and compared with gradient descent (standard back-propagation) for on-line learning in two-layer neural networks with an arbitrary number of hidden units. Within a statistical mechanics framework, both numerical studies and a rigorous analysis show that the adaptive back-propagation method results in faster training by breaking the symmetry between hidden units more efficiently and by providing faster convergence to optimal generalization than gradient descent.

On-line learning with adaptive back-propagation in two-layer networks

An adaptive back-propagation algorithm parameterized by an inverse temperature 1/T is studied and compared with gradient descent (standard back-propagation) for on-line learning in two-layer neural networks with an arbitrary number of hidden units. Within a statistical mechanics framework, we analyse these learning algorithms in both the symmetric and the convergence phase for finite learning rates in the case of uncorrelated teachers of similar but arbitrary length T. These analyses show that adaptive back-propagation results generally in faster training by breaking the symmetry between hidden units more efficiently and by providing faster convergence to optimal generalization than gradient descent.