We consider frequency selective channel estimation in the uplink of massive MIMO-OFDM systems, where our major concern is complexity. A low complexity distributed LMMSE algorithm is proposed that attains near optimal channel impulse response (CIR) estimates from noisy observations at receive antenna array. In proposed method, every antenna estimates the CIRs of its neighborhood followed by recursive sharing of estimates with immediate neighbors. At each step, every antenna calculates the weighted average of shared estimates which converges to near optimal LMMSE solution. The simulation results validate the near optimal performance of proposed algorithm in terms of mean square error (MSE). © 2015 EURASIP.
|Original language||English (US)|
|Title of host publication||2015 23rd European Signal Processing Conference (EUSIPCO)|
|Publisher||Institute of Electrical and Electronics Engineers (IEEE)|
|Number of pages||5|
|State||Published - Dec 28 2015|