Neural Network Based LTE Throughput Estimation in Two-Tier Femtocell Networks

Femtocells are a promising technology that benefits both mobile user and operators by providing economical voice and high speed data service particularly at indoor. Femtocells overlaid on existing macrocell forms a two-tier network and it may be considered as a heterogeneous network. Close access mode working of Femtocell Access Points (FAPs) and sharing radio resource allocation between existing macrocell and femtocells causes both cross-tier and co-tier interference. Achieved user throughput estimation at user equipment and FAP is a key parameter to implement an efficient interference mitigation scheme in the said networks. A neural network based SINR-toChannel Quality Index (CQI) mapping and the throughput estimation scheme is proposed in this work. Femtocell user SINR and throughput estimated using a Neuro-controller estimated CQI are compared with theoretical and actual values. The simulation results obtained verifies the validity of the proposed throughput estimation scheme.