Error probability estimator structures based on analysis of the receiver decision variable

The paper presents schemes for rapid on-line error probability estimation of digital communications links. Several estimator structures are proposed based on the assumption of sample independence, including weighted least squares (WLS) and maximum likelihood (ML) forms. The continuous-form ML estimator is shown to lie on the Rao-Cramer bound, making it a most efficient estimator of probability of error. The design, performance, implementation complexity and behavior of these estimators is described for AWGN