A joint probability density distribution of the instantaneous value
X and the slope
Y in the arbitrary random wave fluctuating only in positive region can be expressed in the form of a mixed orthonormal expansion series in the statistical Laguerre series and the statistical Hermite series as follows:
P(X, Y)=Xm-1/Γ(m)Sme-X/S·1/√2πλe-(Y-μY)2/2λ{1+∞Σn=1∞ΣK=1β(n, K)Ln(m-1)(X/S)HK(Y-μY/√λ)}.
It must be noticed that the general correlation of high degree between level
X and slope
Y is reflected in each expansion coefficient β(
n, k) (
n≠0,
k≠0).
While, the expected number
M(X) of
X per second passing through a level
X with positive and negative slopes is given by
M(X)=∫∞-∞|Y|P(X, Y)dY.
Therefore, by using an expression of the expected number in the normalized form:
M0(X)=M(x)/(∫∞0M(X)dX), the probability density distribution function of crossing at level
X with positive and negative slopes can be expressed as
M0(Z)=Pr(Z;m)+AmPr(1)(Z;m+1)+Bm(m+1)/2!Pr(2)(Z;m+2)+Cm(m+1)(m+2)/3!Pr(3)Z;m+3)+Dm(m+1)(m+2)(m+3)/4!Pr(4)Z;m+4……, where
Z=X/S, S=μx/m, m=μx2/δx2 (μ
x: mean, δ
x2: variance) and
Pr(Z;m)≡Zm-1/Γ(m)e-Z is a gamma probability density distribution function. In the same way, we can obtain an expression of the probability density distribution function
N0(Z) of crossing at level
X with a positive slope in the similar expansion form.
Finally, the detailed experimental considerations to corroborate the above theories are given in the following two cases:
(a) relation between the cumulative probability distribution
∫Z0P(Z)dZ of the instantaneous value and the cumulative probability distribution
∫Z0M0(Z)dZ(or∫Z0N0(Z)dZ) of crossing, (b) high correlations β(
n, k) between level
X and slope
Y, by taking the street noise measured in the city of Fukui as an example of the above random wave fluctuating only in positive region.
View full abstract