I feel my question is somewhat related to this, but I'm not sure how to do what I want.

I get transmissions of ~1000 bits at a time, which may be corrupted by noise. To try and infer data from partially corrupted data, I wish to try and correlate the sequence with expected patterns.

For instance I expect to find a certain subsequence in the stream, followed by a known amount of bits which hold no known structure, and then another set of expected bits. To illustrate:

1010111---n bits unexpected structure - 111000111

In the above, the two specific sequences are expected, and sometimes even the amount of unknown bits in between (n) is apriori known.

The above example is uncorrupted, but if one several bits flip, I wish to scan the stream and look for the best matching pattern.

I already built "masks" of expected patterns and I need to somehow correlate the received data against them, but I'm new to the subject and I'm not sure how to continue from here.

Since I'm new to the subject any: suggestions, references, or proposed schemes will be appreciated.

  • $\begingroup$ have you looked at Hamming distance? $\endgroup$ – user28715 Aug 25 '19 at 16:48
  • $\begingroup$ Thanks, I’m aware of hamming distance. What do you suggest though? Take all possible matches and choose the best one? What cost function would you suggest? $\endgroup$ – Yair M Aug 25 '19 at 16:51
  • 1
    $\begingroup$ cost function? Hamming distance $\endgroup$ – user28715 Aug 25 '19 at 16:52
  • $\begingroup$ Sounds good, how do I formulate it. Do I consider this as searching a tree? Do you have a link or another refrecnce? $\endgroup$ – Yair M Aug 25 '19 at 16:57
  • $\begingroup$ at this point, you are asking a programming question. $\endgroup$ – user28715 Aug 25 '19 at 17:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.