Mode-stirred reverberation chambers: A paradigm for spatio-temporal complexity in dynamic electromagnetic environments

Reverberation chambers a la carte: An overview of the different mode-stirring techniques

“…All these (apparently) different stirring processes are essentially one, from the statistical physics point of view, if they individually satisfy the ergodic hypothesis. A substantial overview of RCs standardized applications and state-of-the-art is sketched in [49].…”

Elliptic Stochastic Fields in Reverberation Chambers

“…The sample noncircularity at this frequency (r S 2 1 ,S * 2 1 = 0.505 + j0.055) causes considerable deviation of the elliptic ensemble (Bessel I) CCDF from the asymptotic circular (χ 2 2 ) CCDF, but is close to the circular sampling Bessel K CCDF for N = 9.8; see below. Significantly improved agreement 6 with the empirical CCDF is obtained by the elliptic sample CCDF (17) for N = 9.8. While excellent agreement exists in the "bulk" central part of the distribution (0 ≤ u/σ U < 2), some discrepancy remains in the intermediate region 2 ≤ u/σ U < 6, where the heavier empirical tail suggests that some uncertainty remains unaccounted for, possibly because the distribution of N has not yet been included in the model (17), or because any statistical dependence between certain distribution parameters may be too strong.…”

“…In [17], it was demonstrated that such sampling PDFs-despite corresponding to the simplest (viz., circular normal) PDF for ensemble E-can yield closer agreement with the empirical distribution of |E| 2 , compared to the agreement with Bessel I ensemble distributions, even though the latter originate from a more sophisticated complex elliptic normal PDF of E. Whether a Bessel K (sampling) PDF for circular E offers better agreement with an empirical PDF than a (ensemble) Bessel I PDF for elliptic E for a particular dataset depends on 1) the NDoF because a relatively low value allows for matching heavier tails, and 2) | E | being not too close to 1, in order that the model ensemble PDF is not oversimplified. Indeed, [17,Fig. 3] has shown that these different aspects may lead to better agreement with either Bessel K PDFs (for |E x | 2 and |E y | 2 ) or Bessel I PDFs (for |E z | 2 ).…”

Source Separation for Wideband Energy Emissions Using Complex Independent Component Analysis

“…Extraction of ICs is also fundamental in the rigorous stochastic characterization of random EM fields in static or dynamic systems when the number of degrees of freedom (independent samples) plays a crucial role [7]. This occurs, e.g., in sampling distributions or in extreme value distributions and their associated statistics and uncertainties for PCB emissions, reverberation chambers, etc.…”

“…For practical reasons explained in Section IV-A, it is further assumed that P n and hence φ n are real valued, leading to a 1-D f P n (p n ) per source. For ideal 6 random EM fields produced by a set of s n -dimensional incoherent circular source currents, such P n exhibit a χ 2 2s n PDF [7]: after standardization of P n to Z P n ,…”