Hi,

I am an Mechanical student and i am trying to generate a random rough

surface (with specified ACF and Std Deviation) using 2D FIR filter in

Matlab. I come across a paper that helps me doing that.

For that, first i need to generate a input sequence composed of

independent random numbers {I,J)}. The paper says its PSD must be a

constant.

I used matlab to generate independent random numbers using the function

randn(94) and tried to find the PSD of the sequence by using the

function 'fft' in matlab. But its not giving me a constant.

Somebody please explain me 1. Why PSD is constant? 2. How to find PSD

using MATLAB?

# PSD of Random Numbers using MATLAB

Started by ●November 4, 2008

Reply by ●November 30, 20082008-11-30

Hi Karthikeyan,

Before concerning about how to find PSD in matlab, you need to make sure that you theoretically know how to find the PSD of a random process. you need to first find the autocorrelation function of your random variable. then taking the "FFT" of that will give you the PSD. this isknown asthe Wiener-Kitchenor recently Wiener-Kitchen-Einstein theorem.

On the other hand, constant PSD means uniform distribution whichmeans any of the random values has equal probability of occurrenceaswiththe others.

I recommend you to do some reading on RandomProcesses and Statistical Analysis.

good luck..

TayyaR.

--- On Mon, 11/3/08, Karthikeyan Kittappan wrote:

From: Karthikeyan Kittappan

Subject: [matlab] PSD of Random Numbers using MATLAB

To: m...

Date: Monday, November 3, 2008, 9:08 PM

Hi,

I am an Mechanical student and i am trying to generate a random rough

surface (with specified ACF and Std Deviation) using 2D FIR filter in

Matlab. I come across a paper that helps me doing that.

For that, first i need to generate a input sequence composed of

independent random numbers {I,J)}. The paper says its PSD must be a

constant.

I used matlab to generate independent random numbers using the function

randn(94) and tried to find the PSD of the sequence by using the

function 'fft' in matlab. But its not giving me a constant.

Somebody please explain me 1. Why PSD is constant? 2. How to find PSD

using MATLAB?

Before concerning about how to find PSD in matlab, you need to make sure that you theoretically know how to find the PSD of a random process. you need to first find the autocorrelation function of your random variable. then taking the "FFT" of that will give you the PSD. this isknown asthe Wiener-Kitchenor recently Wiener-Kitchen-Einstein theorem.

On the other hand, constant PSD means uniform distribution whichmeans any of the random values has equal probability of occurrenceaswiththe others.

I recommend you to do some reading on RandomProcesses and Statistical Analysis.

good luck..

TayyaR.

--- On Mon, 11/3/08, Karthikeyan Kittappan wrote:

From: Karthikeyan Kittappan

Subject: [matlab] PSD of Random Numbers using MATLAB

To: m...

Date: Monday, November 3, 2008, 9:08 PM

Hi,

I am an Mechanical student and i am trying to generate a random rough

surface (with specified ACF and Std Deviation) using 2D FIR filter in

Matlab. I come across a paper that helps me doing that.

For that, first i need to generate a input sequence composed of

independent random numbers {I,J)}. The paper says its PSD must be a

constant.

I used matlab to generate independent random numbers using the function

randn(94) and tried to find the PSD of the sequence by using the

function 'fft' in matlab. But its not giving me a constant.

Somebody please explain me 1. Why PSD is constant? 2. How to find PSD

using MATLAB?