Unsupervised Bit Error Rate Estimation Using Maximum Likelihood Kernel Methods

Abstract

International audienceThis paper addresses the problem of unsupervised BER estimation for communication systems where no prior information about transmitted bits is available at the receiver. We introduce a new technique using kernel pdf estimation together with maximum likelihood-based smoothing parameter computation. Performance is evaluated in the case of CDMA and turbo codes. We show that the new technique provides good BER estimates even in the region of high SNR while it requires a few number of observations

Non-parametric kernel-based bit error probability estimation in digital communication systems: An estimator for soft coded QAM BER computation

“…without requiring the transmitted data to be known. Always based on the Gaussian kernel function, a maximum likelihoodbased smoothing parameter optimisation was studied [13] for the kernel estimators of BEP. Illustrations of the effectiveness of the maximum likelihood-based smoothing parameter selection were made for binary coded transmission schemes involving Turbo and LDPC codes over CDMA systems.…”