The Key Technologies Of Distributed Space-Time Coded Relay Networks

Cooperative communications is a novel technology which can achieve spatial diversity with geographical dispersed wireless nodes equipped with only one antenna. It provides a solution to the problem that it is difficult to set multi-antenna on one node because of the limited size, maximum power and cost. Contrast to traditional communication scheme, cooperative communication can achieve lots of advantages, such as higher reliability, higher power-efficiency, and higher spectrum-efficiency. Recently it is also adopted by industry standards such as WiMAX, LTE etc. Since there are always messes of node in wireless networks, cooperative communication usually appears as distributed space-time (DST) coded relay system. Therefore, it is important to understand and overcome the drawbacks of DST coded relay networks.The main idea of this dissertation is to understand the drawbacks of DST coded relay networks under different relay functions, and to provide useful solutions. Firstly, drawbacks of Amplify-and-forward (AF), Decode-and-Forward (DF), and Demodulation-and-Forward (DmF) are analyzed and listed. Then we also provide the key problems required to be solved in DST coded relay system.DF based DST coded relay system has little differences from multi-antenna system at PHY-layer. Thus cross-layer design and resource allocation are main tasks. For cross-layer design, ARQ protocols for multi-user cooperative communication system, which are post-cooperation ARQ protocol and pre-cooperation ARQ protocol, are designed. Both theoretic and simulation results show that proposed ARQ protocols can improve the system throughput greatly. Considering the running power of relay in resource-limited networks, a power-efficient selective relaying scheme is proposed. Simulation shows that the proposed scheme can improve the system power-efficiency.There are two main problems in AF based DST coded relay systems:the fixed number of relay set limits the ability of supplying different link quality for application-layer service; if the receiver has no idea about channel information, the received noise is not Gaussian any more. For the first problem, a novel distributed space-time coding (DSTC) scheme based on constellation precoding (CP-DSTC) is proposed, that can achieve full-diversity or full-rate, and also offers a flexibility for a desired rate-diversity tradeoff. This code works well with arbitrary signal constellation and any number of relays and achieves minimal-delay. Coding criteria, Chernoff bound, decoding strategies and optimal power allocation are also provided. Simulation results show that proposed code outperforms existing codes at high power regime. In the noise-propagated networks, two-pilot based channel estimation scheme is proposed. Through two pilots, the proposed scheme converts a channel estimation problem in non-Gaussian noise into two channel estimation problems in Gaussian noise. Optimal pilot design and optimal power allocation are both provided. Simulation shows that the proposed channel estimation based receiver can achieve nearly the same performance with perfect channel based one and far outperforms existing channel estimation algorithms.Error propagation is the major problem of DmF based DST coded relay systems. Two SNR thresholds based relay selecting schemes, centralized selecting and distributed selecting, are provided to mitigate error propagation. A novel indicator of the system performance, called signal-error-noise ratio (SENR), is defined herein. The centralized relay selecting scheme is designed to maximize the SENR. For reducing the computational complexity, the distributed selecting scheme is also proposed to minimize the probability of one equivalent relay channel. We also prove that system based on relay selecting can achieve higher diversity order than that with no selecting. Simulation show that proposed relay schemes can efficiently mitigate the error propagation and the centralized selecting with higher complexity can achieve a little improvement than the distributed one.