Performance Evaluation of Optimized Physical Layer Network Coding

With the development of the wireless networks these years, devices are required to have an even higher data rate, and thus the interference between devices remains a significant problem to be solved. In order to secure a reliable transmission with- out wasting network capacity, we would necessarily select a proper transmission scheme. The Physical layer Network Coding (PNC) , is one of the optimum choices as it can increase the efficiency largely, especially in a system with 2 users and 1 relay (TWRC). Generally speaking, we may divide our research into two aspects: channel and channel coding.
From the channel coding’s view, we build the system of uncoded QPSK mod- ulated PNC and convolutional coded PNC (again modulated by QPSK). Due to the feature of the superimposed QPSK constellation, we introduce 3 different fixed mappings (plus the adaptive mapping) in order to decode the superimposed sym- bols at the relay more preciously. Also, we derive the theoretical upper bound of the error rate of the superimposed symbols at the relay. For uncoded PNC, we derive the SER; for coded PNC, we derive the BER, and we focus on an finite accumulation of the terms so as to obtain a tight error bound which is closer to the BER curve of the symbol-level simulation.
From the channel’s view, we build the system of QPSK modulated PNC on single- carrier channel and the PNC on OFDM channels. For PNC on single-carrier channels, we assume the transmission channels are all flat fading channels. Never- theless, it is quite difficult for it to cope with the frequency selective channels and time delay channels. As a result, we introduce OFDM technique which can largely resist their interferences. We develop different adaptive mappings on OFDM inside the sub-band to further improve the performance.
The theoretical upper bound for the error rate of the superimposed signals received at the relay will be vital in the system-level simulation, which is the part of the project DIWINE that my research benefits from.