Why does the Multipath Rayleigh Fading Block produce inaccurate results in certain cases?

2 views (last 30 days)
Why does the Multipath Rayleigh Fading Block produce inaccurate results in certain cases?
I tested the "Communications Blockset/Channels/Multipath Rayleigh Fading Block" with the following parameters:
Doppler frequency: 200
Sample time: 1e-5
Delay vector: 0
Gain vector: 0
Normalization: on
Initial seed: 1234
The input was a real sine wave of 1000 Hz (transformed into a complex input with the "Real-Imag to Complex Block", the imaginary part was a constant zero).
The output was written to the workspace and the Simulation time was 5.3 seconds.
The FFT of the result showed a doppler bandwidth of 2*150 Hz.
Altering the Doppler frequency to 199 Hz showed the same value (2*150 Hz).
Altering the Doppler frequency to 201 Hz showed a Doppler bandwidth of 2*225 Hz.
A few Hz difference could be explained with the necessity of integer interpolation factors, however this difference is too great.

Accepted Answer

MathWorks Support Team
MathWorks Support Team on 27 Jun 2009
By rounding a small real number (for example, 4.5) to an integer, the relative rounding error is quite big, and that makes the Doppler spread inaccurate.
This problem has been fixed in the Communications Blockset version 2.5. If you are using a previous version, one way to solve the problem is to reduce the system's sampling time (e.g., from 1e-5 to 1e-6). The interpolation number will then be 10 times bigger while the rounding error will be much smaller.

More Answers (0)

Categories

Find more on Propagation and Channel Models in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!