Channel Estimation For Two-way Relay Network Based On Network Coding

Most of the existing two-way relay network works (TWRN) have assumed perfect channel state information (CSI) at both the relay node and the two terminals. However, in actual communication system, the channel estimation is required not only for data detection but also for the self-data cancellation at the two terminals. Currently, minor references could be found to discuss the channel estimation of the network coding based TWRN channel. In this thesis, we delve into researching the channel estimation under the network coding situation of the TWRN where two terminals exchange their information through a relay node. A new channel estimation algorithm, denoted as network coding based maximum likelihood (ML) algorithm is proposed.Firstly we study minimum-mean-square-error (MMSE) and vector quantization feedback channel estimation methods: designed for the traditional point-to-point link. However, when they are introduced to the TWRN system, the operation complexity of the relay node as well as the feedback delay and feedback error will be increase.To solve the previous issues, a network coding based maximum likelihood algorithm is proposed for the TWRN system. The basic principle of the algorithm is that the relay amplifies the signal after network coding and broadcasts it to the sources without estimating the channel state information, and then the source nodes directly estimate the complete CSI of the up-down link channel.The thesis derived the specific realization, analyzed the Cramer-Rao lower bound (CRLB) and simulated the performance of the proposed algorithm. The simulation results illustrated that the performance of ML method is always better than least-square (LS) method in the TWRN. With the growth of the transmitter power, the mean square error decreases, and therefore the performance is becoming better.