Fast-Fading Channel Environment Estimation Using Linear Minimum Mean Squares Error-Support Vector Regression

In this article, we propose a Linear Minimum Mean Squares Error-Support Vector Machine regression (LMMSE-SVR) approach which is applied to Long Term Evolution (LTE) downlink channel environment estimation under high mobility conditions. LMMSE-SVR is employed to track and estimate the rapid variations of a realistic Extended Vehicular A model channel according to 3GPP specifications. This contribution assimilates both channel estimation at reference signals and interpolation at data signals into the LMMSE-SVR method. Performances of our channel environment estimation proposal in terms of Bit Error Rate and Mean Squares Error are established via simulation for both normal and extended Cyclic Prefix scenarios in LTE downlink system with 64-QAM modulation scheme under 350 km/h mobile speed.