This paper addresses the problem of channel estimation for MIMO OFDM systems where the channels may be spatially correlated. Starting from OFDM transmission model with independent MIMO channels, we derive a new state-space model that accounts for spatial correlation at transmitter and receiver antenna arrays. Kalman filter is then applied to estimate and track the time-varying channels. Correlation parameters are also estimated from the received data.
Published in:
Statistical Signal Processing, 2003 IEEE Workshop on
Date of Conference: 28 Sept.-1 Oct. 2003