A fast soft bit error rate estimation method

Samir Saoudi (1, 2) , Thomas Derham (3) , Tarik Ait Idir (2, 4, 5) , Patrice Coupé (3)

1 Lab-STICC_TB_CACS_COM
2 SC - Département Signal et Communications
3 Orange Labs [Tokyo]
4 INPT - Institut National des Postes et Télécommunications [Rabat]
5 Lab-STICC - Laboratoire des sciences et techniques de l'information, de la communication et de la connaissance (UMR 3192)

We have suggested in a previous publication a method to estimate the Bit Error Rate (BER) of a digital communications system instead of using the famous Monte Carlo (MC) simulation. This method was based on the estimation of the probability density function (pdf) of soft observed samples. The kernel method was used for the pdf estimation. In this paper, we suggest to use a Gaussian Mixture (GM) model. The Expectation Maximisation algorithm is used to estimate the parameters of this mixture. The optimal number of Gaussians is computed by using Mutual Information Theory. The analytical expression of the BER is therefore simply given by using the different estimated parameters of the Gaussian Mixture. Simulation results are presented to compare the three mentioned methods: Monte Carlo, Kernel and Gaussian Mixture. We analyze the performance of the proposed BER estimator in the framework of a multiuser code division multiple access system and show that attractive performance is achieved compared with conventional MC or Kernel aided techniques. The results show that the GM method can drastically reduce the needed number of samples to estimate the BER in order to reduce the required simulation run-time, even at very low BER.

Domaines: Traitement du signal et de l'image [eess.SP]

Samir Saoudi, Thomas Derham, Tarik Ait Idir, Patrice Coupé. A fast soft bit error rate estimation method. EURASIP Journal on Wireless Communications and Networking, 2010, 2010, pp.372370. ⟨10.1155/2010/372370⟩. ⟨hal-00565746⟩