TY - GEN
T1 - LLR based iterative reduced-complexity MLD algorithm with Parallel Interference Cancellation (PIC) in MIMO systems
AU - Honjo, Katsunari
AU - Ohtsuki, Tomoaki
PY - 2007/8/2
Y1 - 2007/8/2
N2 - We propose a log likelihood ratio (LLR) based iterative reduced-complexity maximum likelihood detection (MLD) algorithm with parallel interference cancellation (PIC) to improve detection performance in multiple-input multiple-output (MIMO) systems. In the proposed algorithm, we estimate the initial channel responses with pilot symbols and detect the initial transmitted symbols by list detection algorithm that is the reduced-complexity suboptimal detection algorithm with detection performance close to maximum likelihood detection (MLD). In the proposed algorithm, we update the channel estimates based on the minimum LLR absolute value given by turbo decoder output only when it is larger than a threshold. Using the updated channel estimates and the transmitted symbol replica, we detect the transmitted symbols by MLD algorithm after performing PIC. In the iterative detection process, we reduce the complexity by reducing the number of candidate symbols of MLD and determine the detection ordering based on LLR. We consider the proposed algorithm in an bit-interleaved turbo coded MIMO system. Simulation results show that the frame error rate (FER) performance of the initial list detection algorithm is improved by performing iterative reduced complexity MLD with PIC.
AB - We propose a log likelihood ratio (LLR) based iterative reduced-complexity maximum likelihood detection (MLD) algorithm with parallel interference cancellation (PIC) to improve detection performance in multiple-input multiple-output (MIMO) systems. In the proposed algorithm, we estimate the initial channel responses with pilot symbols and detect the initial transmitted symbols by list detection algorithm that is the reduced-complexity suboptimal detection algorithm with detection performance close to maximum likelihood detection (MLD). In the proposed algorithm, we update the channel estimates based on the minimum LLR absolute value given by turbo decoder output only when it is larger than a threshold. Using the updated channel estimates and the transmitted symbol replica, we detect the transmitted symbols by MLD algorithm after performing PIC. In the iterative detection process, we reduce the complexity by reducing the number of candidate symbols of MLD and determine the detection ordering based on LLR. We consider the proposed algorithm in an bit-interleaved turbo coded MIMO system. Simulation results show that the frame error rate (FER) performance of the initial list detection algorithm is improved by performing iterative reduced complexity MLD with PIC.
UR - http://www.scopus.com/inward/record.url?scp=34547313506&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=34547313506&partnerID=8YFLogxK
U2 - 10.1109/VETECS.2007.480
DO - 10.1109/VETECS.2007.480
M3 - Conference contribution
AN - SCOPUS:34547313506
SN - 1424402662
SN - 9781424402663
T3 - IEEE Vehicular Technology Conference
SP - 2324
EP - 2328
BT - 2007 IEEE 65th Vehicular Technology Conference - VTC2007-Spring
T2 - 2007 IEEE 65th Vehicular Technology Conference - VTC2007-Spring
Y2 - 22 April 2007 through 25 April 2007
ER -