Collaborative Coding Techniques with Analog Network Coding in Wireless Y-Channel-Relay Networks

摘要

After channel coding reached near Shannon-limit performance with the introduction of Turbo codes and Low-Density Parity-Check (LDPC) codes, research moved on to network coding techniques to enhance overall network performance. The most recent and novel of those approaches is the idea of Physical (Analog) Network Coding which embraces interference from other users, mixes signals in the channel rather than in a relay, and can theoretically increase throughput in the two-way relay channel by up to two folds. In this thesis, we explore this idea, and analyze the theoretical gains of using network coding in a Y-Channel problem - where three users communicate through a common relay. We study existing collaborative coding techniques for the Y-Channel like nested recursive convolutional codes, and Combined Network Channel (CNC). After that, we introduce enhanced nested codes based on turbo codes that achieve good performance in poor SINR conditions. In addition, we propose a novel equal-rate collaborative coding scheme based on algebraic linear block codes. This scheme is simpler to implement than CNC, yet its burst-traffic performance is better than any of the studied solutions. In theory, this code reduces the number of transmission timeslots by up to three folds. Finally, we put forward practical scenarios where physical network coding can be harnessed – mainly in Long Term Evolution (LTE) Multicast (eMBMS), and opportunistic routing in Wireless Mesh Networks.