Abstract:
Different from the orthogonal frequency division multiplexing (OFDM), the filter bank multi-carrier (FBMC) technology, as one of the 5G candidate technologies, involves the design of prototype filter design, so it is unnecessary for each carrier to be completely orthogonal and to insert a cyclic prefix for protection and other operations. However, due to its inherent imaginary interference, FBMC technique can increase the difficulties in channel estimation. Hence, a robust channel estimation and interpolation algorithm is essential to adequately track the time-varying channel response. In this paper, a SVR interpolation-based time-varying channel estimation of FBMC is proposed, in which the auxiliary pilots are utilized to eliminate the imaginary interference. And then, SVR is utilized to interpolate the time-varying channel vector. At the presence of transmission fading and non-linear noise, the proposed algorithm estimates the channel response coefficient by using SVR regression to fit the channel vector. In the simulation, the channel vector is modeled as a single-path slow-time-varying channel with Rayleigh fading. Simulation results show that, compared with the traditional algorithms, the proposed method can greatly improve the accuracy of channel estimation with little loss of efficiency.