Scribd Logo
Upload

Welcome to Scribd!

  • 33 views

    Gaussian and White Random Processes: y X Xy XX Yy XX

    Uploaded by

    Arin Kudlacek
    The document discusses power spectra of linear time-invariant systems and properties of Gaussian and white random processes. It shows that the power spectrum of the output of an LTI system is equal to the power spectrum of the input multiplied by the transfer function. Thermal noise in electronics is modeled as a zero-mean Gaussian process with a flat power spectral density that approximates white noise. The document then discusses band-limited noise processes and decomposing noise into in-phase and quadrature components. Finally, it provides an overview of digital modulation over an additive white Gaussian noise channel.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Gaussian and White Random Processes: y X Xy XX Yy XX

    Uploaded by

    Arin Kudlacek
    33 views4 pages
    The document discusses power spectra of linear time-invariant systems and properties of Gaussian and white random processes. It shows that the power spectrum of the output of an LTI system is equal to the power spectrum of the input multiplied by the transfer function. Thermal noise in electronics is modeled as a zero-mean Gaussian process with a flat power spectral density that approximates white noise. The document then discusses band-limited noise processes and decomposing noise into in-phase and quadrature components. Finally, it provides an overview of digital modulation over an additive white Gaussian noise channel.

    Original Description:

    digital communication

    Original Title

    Lecture 09

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document discusses power spectra of linear time-invariant systems and properties of Gaussian and white random processes. It shows that the power spectrum of the output of an LTI system is equal to the power spectrum of the input multiplied by the transfer function. Thermal noise in electronics is modeled as a zero-mean Gaussian process with a flat power spectral density that approximates white noise. The document then discusses band-limited noise processes and decomposing noise into in-phase and quadrature components. Finally, it provides an overview of digital modulation over an additive white Gaussian noise channel.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    33 views4 pages

    Gaussian and White Random Processes: y X Xy XX Yy XX

    Uploaded by

    Arin Kudlacek
    The document discusses power spectra of linear time-invariant systems and properties of Gaussian and white random processes. It shows that the power spectrum of the output of an LTI system is equal to the power spectrum of the input multiplied by the transfer function. Thermal noise in electronics is modeled as a zero-mean Gaussian process with a flat power spectral density that approximates white noise. The document then discusses band-limited noise processes and decomposing noise into in-phase and quadrature components. Finally, it provides an overview of digital modulation over an additive white Gaussian noise channel.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 4

    Lecture #09 ELEC 464/864 Digital Communication Systems Spring 2015

    Power Spectra of Linear Time-Invariant Systems


    Consider an LTI system where the input/output relationship is given by the
    convolution
    y(t) = h(t) ? x(t).
    It can be shown that

    h(t)dt

    y = x

    Rxy ( ) = Rxx ( ) ? h( )
    Ryy ( ) = Rxx ( ) ? h( ) ? h( ).
    Thus, in the frequency domain we have
    y = x H(0)
    F{h( )}=H (f )

    z }| {
    Sxy (f ) = Sxx (f ) H (f )

    and

    Syx (f ) = Sxx (f )H(f )

    Syy (f ) = Sxx (f )|H(f )|2

    Gaussian and White Random Processes


    Thermal noise generated in resistors and semiconductors is modeled as a
    zero-mean, stationary Gaussian random process N (t) with PSD that is flat
    over a wide range of frequencies.
    Why Gaussian?
    Thermal noise causes many electrons to move in random directions.
    Sum the tiny currents Central Limit Theorem Gaussian
    The PSD of thermal noise is approximated as
    Snn (f ) =

    kT
    .
    2

    where k = 1.38 1023 joules/kelvin is Boltzmans constant and T is the


    temperature in kelvin (cool electronics = low noise). This PSD describes a
    white process, i.e., it contains equal amounts of all frequencies.
    What is wrong with this approximation?
    A more accurate approximation is
    Snn (f ) =

    ~f
    

    2 e
    1

    ~f
    kT

    February 6, 2015

    Lecture #09 ELEC 464/864 Digital Communication Systems Spring 2015

    where ~ is Plancks constant (6.6 1034 Jsec).


    Max at f = 0: kT
    2
    At T = 300 K and f = 2 1012 : Snn (f ) = 0.9 kT
    2
    Thus, the approximation is accurate for frequencies used in traditional wireless communications.
    The PSD of white noise is commonly labelled as
    Snn (f ) =

    N0
    2

    from to , where N0 = kT . This is sometimes referred to as the


    two-sided PSD.
    PSD and autocorrelation:
    Snn (f )

    Rnn ( )

    N0
    2

    N0
    2

    where Rnn ( ) = N20 ( ).


    N (t) and N (t + ) are independent for all 6= 0. If N (t) is the input to
    an LTI system with impulse response h(t), then the system output y(t) is a
    Gaussian random process.
    Band-Limited Noise Processes
    Consider an ideal bandpass filter centered around some frequency fc .
    (
    1 |f fc | W
    H(f ) =
    0 otherwise
    The PSD of the filtered noise is given by
    (
    Sxx (f ) = Snn (f )|H(f )|2 =

    N0
    2

    |f fc | W
    otherwise

    From this, we can calculate the amount of noise present in a band-limited


    signal.

    February 6, 2015

    Lecture #09 ELEC 464/864 Digital Communication Systems Spring 2015

    We can also decompose x(t) into its quadrature and in-phase components
    x(t) = xI cos(2fc t) xQ sin(2fc t),
    where
    xQ and xI are zero-mean, lowpass, jointly WSS, jointly Gaussian processes.
    R
    Total power Px = Sx (f )df = PxQ = PxI
    Understanding of xI and xQ is very important to analyzing digital
    communication schemes like quadrature amplitude modulation (QAM).

    Chapter 8: Digital Modulation in an Additive White


    Gaussian Noise (AWGN) Baseband Channel
    AWGN channel: One of the simplest models for wireless and wired communications.
    AWGN = analog channel We now consider transmitting analog waveforms.
    In this chapter, well assume:
    Baseband transmission
    One-shot communications (no intersymbol interference (ISI))
    x[n]
    Source

    Modulator

    m(t)

    AWGN Channel

    r(t) = m(t) + n(t)


    Receiver
    Well begin by considering the modulator:
    3

    February 6, 2015

    Lecture #09 ELEC 464/864 Digital Communication Systems Spring 2015

    Binary Modulation: x[n] = {0, 1} and modulator transmits s1 (t) for 0


    and s2 (t) for 1.
    M -ary Modulation: Use M = 2k waveforms to transmit k bits at a
    time.
    To simplify analysis, we will develop a vector representation of analog signal
    waveforms.

    February 6, 2015

    You might also like