Issues in the Estimation of Ricean K-Factor from Correlated Samples

Knowledge of the Ricean K-factor is important for accurate assessment and simulation of wireless channels, particularly those in which both terminals are stationary and a strong line-of-sight or fixed component exists. While previous work has focused on development of novel K-factor estimators, the practical matter of determining the minimum number of samples N (or minimum sample duration T) required to estimate K to within a specified confidence interval sigma_K has not received much attention. Here, for the case of a simple moment-based K-factor estimator, we use Monte Carlo simulations to show how N and T depend upon the value of K, the confidence interval sigma_K, the correlation between successive samples, and the sampling interval DeltaT. The results will be useful to those who use such an estimator in practical systems for the purposes of channel state estimation and/or radio resource management.