RAYLEIGH FADING CHANNEL ESTIMATION BASED ON GENERALIZED REGRESSION NEURAL NETWORK

With the rapid development of wireless communication, 5G is gradually growing into a large-scale basic Internet that supports various industries in whole society. The substantial expansion its service scope poses many challenges for underlying technology, especially crucial component physical Layer-Orthogonal Frequency Division Multiplexing (OFDM). Recently, Neural Networks (NNs) have attracted extensive attention due to their excellent performance computing vision and natural language processing. Its strong universality also provides new space traditional communications. This manuscript conducts an in-depth study on channel estimation OFDM systems explores possible application Generalized Regression Network (GRNN) estimate Channel Impulse Response (CIR) attenuated by AWGN Rayleigh fading system. Moreover, three algorithms, i.e., LS, MMSE, LMMSE, are derived mathematics. In addition, this thesis illustrates several typical neural networks detail, including internal structure, parameter updating process, related optimization algorithms.