Physical Layer Network Coding for Mobile Communications

本项目的目标是研究一种新的无线多跳传输方法，它是AhlSwede等人提出的网络编码方案的扩展。在项目过程中，已经开发了多种均衡技术，使这样的物理层网络编码(PNC)中继系统能够在传输受到频率选择性衰落的环境中工作，我们为中继节点设计了三种序列估计方案。这些算法中的两个运行在全状态网格上，因此它们的性能代表了所有随后研究的方法的性能上限。由于基于全状态网格的方案的计算复杂性和内存消耗随着信道冲激响应长度的增加而非常快地变得非常大，我们还设计了一种降低复杂度的方法。然而，复杂性的降低伴随着相当大的性能损失。此外，我们设计了最优的发射滤波方法，允许中继节点使用线性均衡(LE)或判决反馈均衡(DFE)。通过在源节点使用这些滤波器，从两个节点到中继器的传输受到相同总体信道脉冲响应的影响，并且相对于所选择的均衡方法(LE或DFE)的信噪比(SNR)被最大化。一种基于最优Tomlinson-Harashima预编码(THP)的方法补充了上述算法。在源节点采用适当的反馈(预编码)和前馈滤波相结合的方法，可以实现通信节点之间无符号间干扰(ISI)的传输。对于项目的更新，将PNC扩展到单载波频分多址(SC-FDMA)传输将是研究课题之一。SC-FDMA由于其吸引人的特性，已经被标准化用于长期演进(LTE)的上行链路。对于采用SC-FDMA的PNC方案，将研究新的频域均衡器。此外，我们还将考虑更现实的网络拓扑。到目前为止，我们的重点主要是双向中继通道(TWRC)。在该项目的最后一年，我们将把我们的研究扩展到多址中继信道(MARC)。此外，在该项目的工作中，我们还注意到了几个有待进一步研究的问题。基于THP的中继系统目前在广播阶段在目的节点使用DFE。我们将在此阶段开发一种使用THP的方案，以便部分补偿由两个无线信道引起的ISI，并消除具有LE的每个目的节点处的剩余ISI。最后，将开发便于在中继节点使用低复杂度的基于网格的均衡器(如减少状态序列估计(RSSE))的最优发送滤波方案。

The objective of this project is to investigate a novel approach to wireless multi-hop transmission, which is an extension of the network coding scheme as presented by Ahlswede et al. In the course of the project, a variety of equalization techniques that enable such a physical layer network coding (PNC) relaying system to operate in an environment where transmissions are subject to frequency-selective fading have been developed.We have designed three sequence estimation schemes for the relay nodes. Two of these algorithms operate on a full-state trellis and therefore their performance represents an upper bound for the performance of all subsequently investigated approaches. Since the computational complexity and memory consumption of full-state trellis-based schemes becomes prohibitively large very fast as the length of the channel impulse response increases, we also devised a complexity-reduced approach. However, the reduction in complexity comes with a considerable performance penalty. Furthermore, we devised optimal transmit filtering approaches that allow the relay node to either use linear equalization (LE) or decision-feedback equalization (DFE). By employing those filters at the source nodes, the transmissions from both nodes to the relay are influenced by the same overall channel impulse response, and the signal-to-noise ratio (SNR) with respect to the chosen equalization approach (LE or DFE) is maximized. An optimal Tomlinson-Harashima precoding (THP) based approach complements the above mentioned algorithms. Using a combination of suitable feedback (precoding) and feedforward filters at the source nodes, a transmission that is free of intersymbol interference (ISI) can be achieved for both communicating nodes.For the renewal of the project, the extension of PNC to single-carrier frequency-division multiple access (SC-FDMA) transmission will be one of the research topics. SC-FDMA has been already standardized for the uplink of Long Term Evolution (LTE) due to its appealing features. New frequency domain equalizers will be investigated for a PNC scheme with SC-FDMA. Furthermore, we will consider more realistic network topologies. To date, our focus was mainly on the two-way relay channel (TWRC). In the last year of the project, we will expand our investigations to the multiple access relay channel (MARC).Additionally, there are several open points for further research that came to our attention while working on the project. The THP based relaying system currently uses DFE at the destination nodes in the broadcast phase. We will develop a scheme that uses THP instead also in this phase in order to partly compensate for the ISI caused by both radio channels and removes the residual ISI at each destination node with LE. Finally, an optimal transmit filtering scheme that facilitates the use of low complexity trellis-based equalizers like reduced-state sequence estimation (RSSE) at the relay node will be developed.