Channel Estimation based on Echo State Networks in Wireless MIMO Systems

Echo state networks (ESNs) provide architecture and supervised learning principle for recurrent neural networks (RNNs). In this paper, we apply ESN to channel estimation in wireless Multiple-Input and Multiple-Output (MIMO). There is the multipath propagation environment between a transmitter and receiver. Thus the received signal undergoes phase shift, attenuation and time delay. In order to mitigate these random effects and decoding the transmitted signal at the reservoir, we present ESNs for learning nonlinear systems and realizing accurate channel estimation. We also design a teaching scheme to train the output weights of ESNs. The potential for engineering application is illustrated by channel estimation. Numerical results show that accuracy is improved by the number of reservoir units.