Graph simulated not as expected

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #11544
    Fam
    Participant

    Hi,

    The graphs below are the simulations I did using Comparison_CodingSchemes_AWGN_BPSK.m . From the graphs, the only “normal” one is the one with R=1/2, R=1/3 only showing Convolutional Code and Polar Code whereas R=2/3,5/6 and 7/8 are showing uncoded one is better than coded.
    R = 1/3
    R = 1/2
    R = 2/3
    R = 5/6
    R = 7/8

    May I know which other parameter should I take care when I changing the code rate? Because from the graphs I obtained from the simulator is not like what was expected. Those papers I read about channel coding were claiming that LDPC code works better than Turbo code in higher code rate.

    Best Regards,
    Fam

    • This topic was modified 2 years, 10 months ago by Fam.
    • This topic was modified 2 years, 10 months ago by Fam.
    #11552
    Bashar Tahir
    Participant

    Dear Fam,

    The reason is that you are considering the same SNR range for the different code rates. As you increase the code rate, you will need a higher SNR to achieve the same BER. This is because for higher rates, the redundancy is lower, and therefore the error correction capability is worse.

    Also, about the LDPC code working better than turbo; this depends on the constructed codes. There are many LDPC codes and many turbo codes. You have to be specific about it.

    Best,
    Bashar

    • This reply was modified 2 years, 10 months ago by Bashar Tahir.
    • This reply was modified 2 years, 10 months ago by Bashar Tahir.
    #11555
    Bashar Tahir
    Participant

    and you have to specify in which range they work better (higher BER or lower BER), because this also makes a difference, depending on your target application.

    Best,
    Bashar

    #11620
    Fam
    Participant

    Dear Bashar,

    Noted with thanks!

    There is another question here, may I know why is the graph simulated stopped before SNR=5dB?
    For example: R=1/2, graph for turbo and LDPC codes stopped at SNR of 1.5dB while convolutional stopped at SNR of 4.5dB, whereas polar codes manage to get the graph until SNR of 30dB.
    As I set the SNR as followed:

    SNRvecdB = linspace(0, 30, 30);
    BER = zeros(30,length(SNRvecdB));

    It supposed to have more readings until SNR of 30dB right?

    Best Regards,
    Fam

Viewing 4 posts - 1 through 4 (of 4 total)
  • You must be logged in to reply to this topic.