To overcome the drawbacks of samples impoverishment, an evolutionary particle filtering algorithm is proposed for blind signal detection over flat Rayleigh fading channels whose model coefficients are unknown. During the resampling of particle filtering, evolutionary programming is used. The evolutionary particle filtering algorithm is adopted to overcome the drawbacks of samples, which forms into a blind detection based in evolutionary particle filtering. It is shown through simulations that the proposed particle filtering detector maintains comparable performance with mixture Kalman filter with known model coefficients.
Hukumchand K J. Sequential Monte Carlo methods fordynamic state space models with applications tocommunications. [Ph.D. Dissertation] State Uniersity of NewYork, 2001.[2]Djuric P M and Kotecha J H. Particle filtering: a review ofthe theory and how it can be used for solving problems in wireless communications. IEEE Signal Processing Magazine,2003, 20(5): 19-38.[3]Chen R, Wang X, and Liu J S. Adaptive joint detection anddecoding in flat-fading channels via mixture Kalman filtering[J].IEEE Trans. on Information Theory.2000, 46(6):2079-2094[4]Punskaya E, Andrieu C, and Doucet A. Particle filtering fordemodulation in fading channels with non-gaussian additivenoise. IEEE Trans. on Commun., 2001, 49(4): 572-582.[5]Huang Y and Djuric P M. A blind particle filtering detectorof signals transmitted over flat fading channels[J].IEEE Trans.on Signal Processing.2004, 52(7):1891-1900[6]Pitt M and Shephard N. Filtering via simulation: auxiliaryparticle filter[J].J. Amer. Statist. Assoc.1999, 94(4):590-599[7]Mo Yi-wei and Xiao De-yun. Evolutionary particle filter andits application. Control Theory & Applications, 2005, 22(2):269-272.