Ergodic radio resource allocation based on imperfect channel distribution information

In this paper, the effect of channel distribution information (CDI) imperfectness on the performance of wireless networks is investigated. In the literature, ergodic resource allocation problems are solved assuming that perfect CDI is available which might not be a practical assumption. Therefore, we adopt the nonparametric density estimation methods for estimating the channel gain distribution. The estimation is first carried out through the well-known kernel density estimation (KDE) method. Since KDE is too sensitive to contaminated data, we adopt the robust kernel density estimation (RKDE) method. The analysis is performed over an ergodic resource allocation problem framework in the uplink of an orthogonal frequency division multiple access based network. In the proposed problem, the objective is to maximize the average total rate subject to total power constraint for each user. Simulation results indicate that for large enough number of nominal data and a reasonable number of outlier data, RKDE can provide a sum rate very close to the one obtained based on the actual CDI.