Semi-Blind Channel Estimation and Symbol Detection Using Combined Superimposed Training
Xianwen He, Gaoming Huang, Gaoqi Dou, Jun Gao · 2016
In amplify-and forward (AF) cooperative systems, accurate channel state information (CSI) for both the cascaded and individual links is essential for coherent combining to achieve spatial diversity. The combined superimposed training strategy(TM+ST) is proposed, where the time-multiplexed (TM) training scheme is employed at source node (SN) to estimate the cascaded channel(S-R-D) while relay-assisted training scheme is employed to extract the individual channel of R-D link at destination node (DN) without bandwidth expansion. The Data-Dependent Distortion (DDD) removed iteration scheme is presented to improve the estimation performance of R-D link. Compared with the traditional estimation strategy, the new approach is proposed to optimize the design of training, making individual channel and cascaded channel estimated independently, which improves the flexibility of the system design. Simulation results are presented that diversity combination can eliminate the symbol error floor and assess the performance of the proposed schemes.