Blind estimation of channel BERs in a multi-receiver network

Abstract

In wireless communications, it is often desirable to merge bit decisions from multiple receivers to improve overall link performance. It is well known that in order to optimally fuse bit decisions from a network of receivers, precise knowledge of receiver bit error rates (BERs) is needed. This information, however, is rarely available in practice. In this work, we present an iterative procedure for blindly estimating receivers BERs to enable near optimal blind fusion of bit decisions in a multi-receiver network. We show that the solution of the estimation problem is a structured eigenvalue task and propose a modified power method procedure to perform it. We prove that the desired solution is a stable point of the algorithm and the algorithm is locally stable. Furthermore, we show via simulations that the technique results in excellent performance in nearly all practical operating scenarios.

Performance Analysis of Optimal Blind Fusion of Bits

“…In [4]- [6], it is claimed that these approaches are optimal and this assertion is supported by two points. The statistic carries the same information about as the observation , and the random variables are independent.…”

“…5. Probability of error given by the "optimal" fusion (2.2) derived from channel BER estimated by the Dixit (4.2) (o), [6] (+), and the EM (*) algorithms respectively, compared with the "optimal" fusion (-) derived from the exact values of ( ) as a function of the sample size.…”

“…Indeed in the case of equality of 0 and 1, the majority detector [6] (+) and the EM algorithm (*) of the parameter as a function of the sample size. (b) Probability of error given by the "optimal" fusion (2.2) derived from channel BER estimated by the Dixit rule (4.2) (o), [6] (+) and the EM (*) algorithms respectively, compared with the "optimal" fusion (-) derived from the exact values of ( ) as a function of the sample size.…”

“…Assuming perfect knowledge of the individual BERs of each detector, optimal fusion of bit estimates is a standard problem [3]. However, when this knowledge is missing, only recent empirical algorithms [4]- [6] have addressed this problem. Reference [4] introduces the problem of blind fusion of bit estimates, curiously not studied until now: It presents a very clever, simple, and intuitive algorithm to solve this problem.…”

“…Reference [4] introduces the problem of blind fusion of bit estimates, curiously not studied until now: It presents a very clever, simple, and intuitive algorithm to solve this problem. Then, a refinement of the previous algorithm and an iterative procedure that outperforms the previous algorithms for small numbers of observations are proposed in [5] and [6], respectively. In this contribution, we examine the statistical performance of blindly estimating bit error rates (BERs).…”