Rateless Codes And Their Applications In Relay Systems

As the research focus of wireless communication technology in recent years, rateless codes and cooperative relaying have been proved to achieve higher reliability, efficiency and robustness for wireless communication systems. On one hand, for rateless codes, their characteristics of forward incremental redundancy, adaptive transmission and little feedback, make them attractive since the transmitter can adaptively adjust the coding rate to fit the channel capacity without retransmission, which not only simplifies the design of system signaling, but also improves the system reliability and throughput simultaneously. Thus, designing a class of rateless codes with good performance has become the key of the research field of rateless codes. On the other hand, cooperative relaying performs significant improvement in data rate, spectral efficiency and energy efficiency. It is desired to achieve further improvement of performance if rateless codes are integrated with cooperative relaying. In this dissertation, we mainly investigate on the design of rateless codes and their applications in two-way relay systems and buffered-relay systems, and propose some novel coding design approaches and rateless coded transmission protocols. The contents of this dissertation are listed as follows.We develop a method for the design of complexity-optimized Raptor codes, which are one class of the classic rateless codes, for the binary input additive white Gaussian noise (BIAWGN) channel under the joint belief propagation (BP) decoder. Since the degree distribution plays a pivotal role in the performance of Raptor codes including rate performance and decoding com-plexity, we aim to obtain an optimal degree distribution minimizing the decoding complexity of Raptor codes without much loss of rate performance. Utilizing the BP decoder, the decoding complexity is measured by the average number of arithmetic operations needed to correctly re-cover each information bit, which shows that the decoding complexity is linear to the number of decoding iterations. Based on the analytical asymptotic convergence analysis which is built upon extrinsic information transfer (EXIT) charts, we develop a numerical approximation for the number of iterations, and formulate an optimization problem for the design of efficient output degree distributions. The proposed optimization problem is to minimize the decoding complexi-ty subject to the convergence constraint, starting condition, stability condition and rate constraint which accounts for rate-complexity tradeoff. Simulation results show that the optimized degree distribution indeed achieves lower complexity without much rate performance loss.In the second part of this dissertation, we propose a novel end-to-end rateless coded physical layer network coding (PNC) scheme for two-way relay systems. In such system, the two-stage relaying strategy is considered, i.e., both end nodes transmit rateless coded symbols at the first stage, and then the relay broadcasts after performing PNC symbol by symbol at the second stage. The two-way relay channel is studied from a new perspective of end-to-end channel equivalence, and a novel end-to-end demodulator is proposed based on the resultant end-to-end equivalent channel with ideal channel side information. Unlike the conventional demodulator for the point-to-point channel, our proposed demodulator takes the link noise and the potential error brought by PNC mapping at the relay into account simultaneously, thus improves the system reliability and throughput. Further, the theoretical throughput is derived by both end-to-end channel equivalence and concatenated channel equivalence based on information theory, which serves as an upper bound of end-to-end achievable rate for the proposed demodulator. Simulation results validate the high efficiency of the proposed scheme in terms of bit error rate (BER) and system throughput compared to the conventional one.Finally, rateless codes are further introduced into half-duplex buffered-relay systems and a novel rateless coded opportunistic relaying protocol is proposed. By employing rateless codes, the transmitters can automatically adapt their coding rate according to the instantaneous channel state information (CSI). Furthermore, the buffered relay provides extra freedom for achieving throughput gain even though the direct link between the source node and destination node is unavailable. Under the proposed opportunistic relaying protocol, the relay would first perform adaptive link selection according to the instantaneous CSI of both links and the queue state of the buffer, and then the selected transmitter sends a rateless coded packet. Utilizing the first-in first-out (FIFO) queueing model, we propose the optimal opportunistic relaying strategies for both scenarios that the buffer size is infinite and finite, and derive the system throughput and end-to-end average delay mathematically. Especially, for the finite buffered-relay system, the buffered queue is modeled by a Markov chain based on the per packet fading model that is widely exploited in rateless coded systems, thus the analysis for the queueing system can be for-mulated exactly. Furthermore, we design an optimal opportunistic relaying protocol maximizing the throughput under the constraint that the end-to-end average delay should meet some practi-cal quality of service (Qos). Simulations verify our analytical results and show the significant throughput improvement compared to the conventional protocol without buffer.