Sadeghi, ParastooRapajic, Predrag2015-12-08May 17 2001090-3038http://hdl.handle.net/1885/29132In this paper, an information theoretic approach is taken to the problem of joint channel estimation and coherent data detection in time varying flat fading channels. At the receiver side, the fading phase process is partitioned or mapped into a finite state Markov channel (FSMC). Using simulation-based capacity analysis of FSMC models, accurate capacity calculation of time varying fading channels is carried out and the effect of number of channel phase partitions on the channel capacity performance is studied. In BPSK scheme, the capacity is saturated beyond 8 to 16 channel phase partitions. In QPSK signalling, the channel capacity gap between 16/32 channel phase partitions and 128 channel phase partitions is less than 0.3 dB. Moreover, the capacity performances of BPSK and QPSK in unknown correlated fading are compared. The comparison reveals that to maintain the superior QPSK capacity performance, care must be taken in the choice of average transmitted power and symbol duration.Keywords: Bit error rate; Channel capacity; Computational complexity; Computer simulation; Mathematical models; Phase locked loops; Phase shift keying; Probability density function; Rayleigh fading; Signal to noise ratio; Channel phase estimation; Fading channels;200410.1109/VETECF.2004.14004222015-12-08