An improved differential phase detection (DPD) scheme based on maximum likelihood sequence estimation (MLSE) is presented for M-ary DPSK modulation. The transmitted phase sequence is estimated from the detected received signal phases only. From the approximate a posteriori joint probability distribution of the output phase noises of L differential phase detectors with orders from 1 to L symbols, the author derives the suboptimal decision algorithm called ML-DPD. To achieve better performance than ML-DPD while also reducing computation complexity, he derives Viterbi-decoding DPD (Viterbi-DPD) which uses, as the branch metric, the sum of squared phase errors of L differential phase detectors with orders from 1 to L symbols. The performance improvements possible with these two new detection schemes are evaluated by computer simulations for 2- and 4 DPSK transmission in AWGN channels.
ASJC Scopus subject areas
- Electrical and Electronic Engineering