Hi all, I am using Viterbi and Viterbi algorithm for estimating phase of the carrier in a 8psk demodulator. I am facing problem relating to cycle slips and observing that half of my burst is demodulated correctly and the rest of the burst has a constant phase offset. My training symbol is able to correct the phase offset in the section before cycle slip. I have tried to find out the cycle slip correction algorithms for this but all I found on ieeexplore is BER analysis and expression for probability of cycle slips. Does there exist any cycle slip correction algorithm for psk demodulators at all? Please guide. Aitezaz
Cycle Slip in Viterbi and Viterbi Non-linear Phase Estimator
Started by ●July 14, 2009
Reply by ●July 14, 20092009-07-14
Aitezaz wrote:> Hi all, > I am using Viterbi and Viterbi algorithm for estimating phase of the > carrier in a 8psk demodulator. I am facing problem relating to cycle > slips and observing that half of my burst is demodulated correctly and > the rest of the burst has a constant phase offset.This tells that your carrier recovery algorithm is not working properly. The cycle slips should not happen until the demodulated data is completely unusable.> My training symbol > is able to correct the phase offset in the section before cycle slip. > I have tried to find out the cycle slip correction algorithms for this > but all I found on ieeexplore is BER analysis and expression for > probability of cycle slips. Does there exist any cycle slip correction > algorithm for psk demodulators at all? Please guide.If your coding is not rotationaly invariant, then you can attempt the decoding for every possible phase and select the phase which has the best metrics. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
Reply by ●July 14, 20092009-07-14
Thanks Vladimir for your reply> This tells that your carrier recovery algorithm is not working properly. > The cycle slips should not happen until the demodulated data is > completely unusable.Yeah. And I face this only on low SNRs. Like this behavior I observe with SNR < 10 and that too randomly and not every time i simulate. On higher SNRs there is no cycle slips. So still you think my carrier recovery has some problem?> If your coding is not rotationaly invariant, then you can attempt the > decoding for every possible phase and select the phase which has the > best metrics.If you can suggest some starting material for rotationaly invariant coding I would be very thankful to you. For your reference, I need to build pragmatic 8TCM. So, if I can solve this problem in the viterbi decoder please suggest. Thanks again for your time.
Reply by ●July 14, 20092009-07-14