1,721,166 research outputs found
Fault detection in dielectric grid scattering
The problem of diagnosing a grid of small (in terms of the
probing wavelength) dielectric scatterers is considered. The aim is to
detect and locate possible defects occurring within a known grid when
one (or more) scatterer is removed/missing (fault). The study is developed
for the canonical case of a TM scalar two-dimensional geometry with
the scatterers consisting of dielectric cylinders of small circular cross
section. The scattering by a fault is modeled by relaying only to a priori
information about the complete grid which leads to a numerically effective
inversion procedures as the bulk of the numerical effort is to be done only
once. Inversion is achieved by a truncated singular value decomposition
scheme and results are provided in terms of closed form expressions for the
probability of detection and of false alarm. This allows us to foreseen the
achievable performance and to highlight the role of scattering configuration
parameters. Numerical examples are also enclosed to corroborate theoretical
outcomes. The case of two or more faults is considered as well. For such
a case it is numerically shown that detection method still works well even
though multiple scattering (occurring between faults) is neglected
Study of unequally-excited random antenna arrays for beam shaping
Random arrays have been typically studied by considering real uniform excitations. This is suited for single-beam radiation patterns but does not allow for more sophisticated patterns. Indeed, only even patterns, with respect to the steering angle, can be achieved. To overcome this limitation, we recently proposed a new model whereby the excitation coefficients are not uniform and are determined by means of two random variable transformations. In this paper, we deal more extensively with the properties of this model, highlighting things that have not been pointed out previously. In order to get analytical results, we just consider symmetric random arrays. For such a case, we determine the design error, that is the cumulative distribution function of the supremum of the the difference between the actual and desired array factors. It is shown that general shaped beams can be actually achieved but at the cost of an increase of the design error as compared to the single-beam case. Numerical analysis validates the presented theory
- …
