Title :
FDD overhead optimization for a multiuser two-way system with imperfect CSI
Author :
Slim, Israa ; Neda, Mario H Casta ; Mezghani, Amine ; Nossek, Josef A.
Author_Institution :
Inst. for Circuit Theor. & Signal Process., Tech. Univ. Munchen, Munich, Germany
Abstract :
In this paper, we find the optimum lengths of TUL, TDL and B i.e. the uplink (UL) and downlink (DL) pilots and feedback bits respectively maximizing the sum of the lower bounds of the UL and the DL rates. With the aid of TUL symbols, the Base Station (BS) estimates the UL channel of each user. TDL symbols are means of estimating the downlink channel, which after being normalized and quantized with B bits, is made available at the BS. Estimated Channel State Information (CSI) and quantized Channel Direction Information (CDI), which may be received erroneously, are thus available at the BS for computing the receive and the transmit Zero-forcing (ZF) filters, respectively. The use of ZF enables the derivation of analytical DL and UL mean squared error (MSE) expressions that give a lower bound on the DL and UL rates, respectively, needed for our optimization.
Keywords :
channel estimation; filtering theory; mean square error methods; multiuser channels; quantisation (signal); radiocommunication; FDD overhead optimization; TDL; TUL; downlink channel; estimated channel state information; feedback bits; imperfect CSI; mean squared error expression; multiuser two way system; quantized channel direction information; transmit zero-forcing filter; uplink channel; Channel estimation; Channel models; Covariance matrix; Estimation error; Optimization; Phase shift keying; Quantization; Broadcast Channel (BC); Estimated Channel State Information; Multiple Access Channel (MAC); Overhead Optimization; Quantized Channel Direction Information; ZF filters; limited Feedback;
Conference_Titel :
Smart Antennas (WSA), 2011 International ITG Workshop on
Conference_Location :
Aachen
Print_ISBN :
978-1-61284-075-8
Electronic_ISBN :
978-1-61284-073-4
DOI :
10.1109/WSA.2011.5741923