Complex random vector
≈ 9 min read
Names in other languages
No names in other languages are held for this entry yet.
Article bodies are English and are not machine-translated. A name in another language is not a translation of the article.
In probability theory and statistics, a complex random vector is typically a tuple of complex-valued random variables, and generally is a random variable taking values in a vector space over the field of complex numbers. If Z1,…,Zn are complex-valued random variables, then the n-tuple (Z1,…,Zn) is a complex random vector. Complex random variables can always be considered as pairs of real random vectors: their real and imaginary parts.
Some concepts of real random vectors have a straightforward generalization to complex random vectors. For example, the definition of the mean of a complex random vector. Other concepts are unique to complex random vectors.
Applications of complex random vectors are found in digital signal processing.
Definition
edit edit sourceA complex random vector 𝐙=(Z1,…,Zn)T on the probability space (Ω,ℱ,P) is a function 𝐙:Ω→ℂn such that the vector (ℜ(Z1),ℑ(Z1),…,ℜ(Zn),ℑ(Zn))T is a real random vector on (Ω,ℱ,P) where ℜ(z) denotes the real part of z and ℑ(z) denotes the imaginary part of z.[1]: p. 292
Cumulative distribution function
edit edit sourceThe generalization of the cumulative distribution function from real to complex random variables is not obvious because expressions of the form P(Z≤1+3i) make no sense. However expressions of the form P(ℜ(Z)≤1,ℑ(Z)≤3) make sense. Therefore, the cumulative distribution function F𝐙:ℂn↦[0,1] of a random vector 𝐙=(Z1,...,Zn)T is defined as
| F𝐙(𝐳)=P(ℜ(Z1)≤ℜ(z1),ℑ(Z1)≤ℑ(z1),…,ℜ(Zn)≤ℜ(zn),ℑ(Zn)≤ℑ(zn)) | Eq.1 |
where 𝐳=(z1,...,zn)T.
Expectation
edit edit sourceAs in the real case the expectation (also called expected value) of a complex random vector is taken component-wise.[1]: p. 293
| E[𝐙]=(E[Z1],…,E[Zn])T | Eq.2 |
Covariance matrix and pseudo-covariance matrix
edit edit sourceThe covariance matrix (also called second central moment) K𝐙𝐙 contains the covariances between all pairs of components. The covariance matrix of an n×1 random vector is an n×n matrix whose (i,j)th element is the covariance between the i th and the j th random variables.[2]: p.372 Unlike in the case of real random variables, the covariance between two random variables involves the complex conjugate of one of the two. Thus the covariance matrix is a Hermitian matrix.[1]: p. 293
|
K𝐙𝐙=cov[𝐙,𝐙]=E[(𝐙−E[𝐙])(𝐙−E[𝐙])H]=E[𝐙𝐙H]−E[𝐙]E[𝐙H] | Eq.3 |
- K𝐙𝐙=[E[(Z1−E[Z1])(Z1−E[Z1])‾]E[(Z1−E[Z1])(Z2−E[Z2])‾]⋯E[(Z1−E[Z1])(Zn−E[Zn])‾]E[(Z2−E[Z2])(Z1−E[Z1])‾]E[(Z2−E[Z2])(Z2−E[Z2])‾]⋯E[(Z2−E[Z2])(Zn−E[Zn])‾]⋮⋮⋱⋮E[(Zn−E[Zn])(Z1−E[Z1])‾]E[(Zn−E[Zn])(Z2−E[Z2])‾]⋯E[(Zn−E[Zn])(Zn−E[Zn])‾]]
The pseudo-covariance matrix (also called relation matrix) is defined replacing Hermitian transposition by transposition in the definition above.
|
J𝐙𝐙=cov[𝐙,𝐙‾]=E[(𝐙−E[𝐙])(𝐙−E[𝐙])T]=E[𝐙𝐙T]−E[𝐙]E[𝐙T] | Eq.4 |
- J𝐙𝐙=[E[(Z1−E[Z1])(Z1−E[Z1])]E[(Z1−E[Z1])(Z2−E[Z2])]⋯E[(Z1−E[Z1])(Zn−E[Zn])]E[(Z2−E[Z2])(Z1−E[Z1])]E[(Z2−E[Z2])(Z2−E[Z2])]⋯E[(Z2−E[Z2])(Zn−E[Zn])]⋮⋮⋱⋮E[(Zn−E[Zn])(Z1−E[Z1])]E[(Zn−E[Zn])(Z2−E[Z2])]⋯E[(Zn−E[Zn])(Zn−E[Zn])]]
- Properties
The covariance matrix is a hermitian matrix, i.e.[1]: p. 293
- K𝐙𝐙H=K𝐙𝐙.
The pseudo-covariance matrix is a symmetric matrix, i.e.
- J𝐙𝐙T=J𝐙𝐙.
The covariance matrix is a positive semidefinite matrix, i.e.
- 𝐚HK𝐙𝐙𝐚≥0for all 𝐚∈ℂn.
Covariance matrices of real and imaginary parts
edit edit sourceBy decomposing the random vector 𝐙 into its real part 𝐗=ℜ(𝐙) and imaginary part 𝐘=ℑ(𝐙) (i.e. 𝐙=𝐗+i𝐘), the pair (𝐗,𝐘) has a covariance matrix of the form:
- [K𝐗𝐗K𝐗𝐘K𝐘𝐗K𝐘𝐘]
The matrices K𝐙𝐙 and J𝐙𝐙 can be related to the covariance matrices of 𝐗 and 𝐘 via the following expressions:
- K𝐗𝐗=E[(𝐗−E[𝐗])(𝐗−E[𝐗])T]=12Re(K𝐙𝐙+J𝐙𝐙)K𝐘𝐘=E[(𝐘−E[𝐘])(𝐘−E[𝐘])T]=12Re(K𝐙𝐙−J𝐙𝐙)K𝐘𝐗=E[(𝐘−E[𝐘])(𝐗−E[𝐗])T]=12Im(J𝐙𝐙+K𝐙𝐙)K𝐗𝐘=E[(𝐗−E[𝐗])(𝐘−E[𝐘])T]=12Im(J𝐙𝐙−K𝐙𝐙)
Conversely:
- K𝐙𝐙=K𝐗𝐗+K𝐘𝐘+i(K𝐘𝐗−K𝐗𝐘)J𝐙𝐙=K𝐗𝐗−K𝐘𝐘+i(K𝐘𝐗+K𝐗𝐘)
Cross-covariance matrix and pseudo-cross-covariance matrix
edit edit sourceThe cross-covariance matrix between two complex random vectors 𝐙,𝐖 is defined as:
| K𝐙𝐖=cov[𝐙,𝐖]=E[(𝐙−E[𝐙])(𝐖−E[𝐖])H]=E[𝐙𝐖H]−E[𝐙]E[𝐖H] | Eq.5 |
- K𝐙𝐖=[E[(Z1−E[Z1])(W1−E[W1])‾]E[(Z1−E[Z1])(W2−E[W2])‾]⋯E[(Z1−E[Z1])(Wn−E[Wn])‾]E[(Z2−E[Z2])(W1−E[W1])‾]E[(Z2−E[Z2])(W2−E[W2])‾]⋯E[(Z2−E[Z2])(Wn−E[Wn])‾]⋮⋮⋱⋮E[(Zn−E[Zn])(W1−E[W1])‾]E[(Zn−E[Zn])(W2−E[W2])‾]⋯E[(Zn−E[Zn])(Wn−E[Wn])‾]]
And the pseudo-cross-covariance matrix is defined as:
| J𝐙𝐖=cov[𝐙,𝐖‾]=E[(𝐙−E[𝐙])(𝐖−E[𝐖])T]=E[𝐙𝐖T]−E[𝐙]E[𝐖T] | Eq.6 |
- J𝐙𝐖=[E[(Z1−E[Z1])(W1−E[W1])]E[(Z1−E[Z1])(W2−E[W2])]⋯E[(Z1−E[Z1])(Wn−E[Wn])]E[(Z2−E[Z2])(W1−E[W1])]E[(Z2−E[Z2])(W2−E[W2])]⋯E[(Z2−E[Z2])(Wn−E[Wn])]⋮⋮⋱⋮E[(Zn−E[Zn])(W1−E[W1])]E[(Zn−E[Zn])(W2−E[W2])]⋯E[(Zn−E[Zn])(Wn−E[Wn])]]
Two complex random vectors 𝐙 and 𝐖 are called uncorrelated if
- K𝐙𝐖=J𝐙𝐖=0.
Independence
edit edit sourceTwo complex random vectors 𝐙=(Z1,...,Zm)T and 𝐖=(W1,...,Wn)T are called independent if
| F𝐙,𝐖(𝐳,𝐰)=F𝐙(𝐳)⋅F𝐖(𝐰)for all 𝐳,𝐰 | Eq.7 |
where F𝐙(𝐳) and F𝐖(𝐰) denote the cumulative distribution functions of 𝐙 and 𝐖 as defined in Eq.1 and F𝐙,𝐖(𝐳,𝐰) denotes their joint cumulative distribution function. Independence of 𝐙 and 𝐖 is often denoted by 𝐙⊥⊥𝐖. Written component-wise, 𝐙 and 𝐖 are called independent if
- FZ1,…,Zm,W1,…,Wn(z1,…,zm,w1,…,wn)=FZ1,…,Zm(z1,…,zm)⋅FW1,…,Wn(w1,…,wn)for all z1,…,zm,w1,…,wn.
Circular symmetry
edit edit sourceA complex random vector 𝐙 is called circularly symmetric if for every deterministic φ∈[−π,π) the distribution of eiφ𝐙 equals the distribution of 𝐙.[3]: pp. 500–501
- Properties
- The expectation of a circularly symmetric complex random vector is either zero or it is not defined.[3]: p. 500
- The pseudo-covariance matrix of a circularly symmetric complex random vector is zero.[3]: p. 584
Proper complex random vectors
edit edit sourceA complex random vector 𝐙 is called proper if the following three conditions are all satisfied:[1]: p. 293
- E[𝐙]=0 (zero mean)
- var[Z1]<∞,…,var[Zn]<∞ (all components have finite variance)
- E[𝐙𝐙T]=0
Two complex random vectors 𝐙,𝐖 are called jointly proper if the composite random vector (Z1,Z2,…,Zm,W1,W2,…,Wn)T is proper.
- Properties
- A complex random vector 𝐙 is proper if, and only if, for all (deterministic) vectors 𝐜∈ℂn the complex random variable 𝐜T𝐙 is proper.[1]: p. 293
- Linear transformations of proper complex random vectors are proper, i.e. if 𝐙 is a proper random vectors with n components and A is a deterministic m×n matrix, then the complex random vector A𝐙 is also proper.[1]: p. 295
- Every circularly symmetric complex random vector with finite variance of all its components is proper.[1]: p. 295
- There are proper complex random vectors that are not circularly symmetric.[1]: p. 504
- A real random vector is proper if and only if it is constant.
- Two jointly proper complex random vectors are uncorrelated if and only if their covariance matrix is zero, i.e. if K𝐙𝐖=0.
Cauchy–Schwarz inequality
edit edit sourceThe Cauchy–Schwarz inequality for complex random vectors is
- |E[𝐙H𝐖]|2≤E[𝐙H𝐙]E[|𝐖H𝐖|].
Characteristic function
edit edit sourceThe characteristic function of a complex random vector 𝐙 with n components is a function ℂn→ℂ defined by:[1]: p. 295
- φ𝐙(𝝎)=E[eiℜ(𝝎H𝐙)]=E[ei(ℜ(ω1)ℜ(Z1)+ℑ(ω1)ℑ(Z1)+⋯+ℜ(ωn)ℜ(Zn)+ℑ(ωn)ℑ(Zn))]
See also
edit edit source- Complex normal distribution
- Complex random variable (scalar case)
References
edit edit source- ^ a b c d e f g h i j Lapidoth, Amos (2009). A Foundation in Digital Communication. Cambridge University Press. ISBN 978-0-521-19395-5.
- ^ Gubner, John A. (2006). Probability and Random Processes for Electrical and Computer Engineers. Cambridge University Press. ISBN 978-0-521-86470-1.
- ^ a b c Tse, David (2005). Fundamentals of Wireless Communication. Cambridge University Press.
Text derived from https://en.wikipedia.org/w/index.php?oldid=1320556974 under CC BY-SA 4.0; modifications have been made.