Study On Two-Way Relaying Protocol And Performance Analysis

Network coding has been proved as an efficient method to improve the throughput and reduce the energy consumption for the wireless two-way relay network. In practical physical transmission, the payload transmission requires a considerable necessary signaling overhead, such as the Physical Layer Convergence Protocol Preamble code, as well as the acquisition of the channel information and the transmission of the scheduling. As a result, it is reasonable and necessary to consider the signaling overhead in the practical two-way relay performance analysis. Unfortunately, all of the current researches on the two-way relay ignored the signaling overhead during the modelling and performance analysis. The buffer resource is a critical resource in the two-way relay network, and the correctly received packets can be stored in the queue and wait for a proper opportunity to be forwarded, this storage is useful to enhance the system performance. However, the most existing paper focused on the optimization problem of the achievable sum-rate, the compromise between the throughput and the queueing delay, as well as the throughput under the constrained delay, and the finite relay has not been fully studied. In addition to that, the transmission character of wireless link cannot always guarantee a successful delivery, while most existing studies assumed an error-free link.To this end, this dissertation investigates the transmission strategies and the associated performance of the two-way relay under the joint consideration of the finite relay buffer, lossy link and non-ignorable signaling overhead. The main contributions of this dissertation can be summarized as follows,a) A two-way relay transmission strategy without scheduling is proposed with the joint consideration of signalling overhead, finite buffer and non-ignorable lossy link. An analytic model is constructed to capture the character of the transmission protocol. Since the transmission of the scheduling message degrades the throughput significantly, the proposed transmission protocol takes no account of the scheduling assignment, which saves a considerable of signalling resources. With the mathematical analysis, we derived the closed form expressions of the system performance, and then gave four Lemmas, which indicate that, the throughput is a monotonic increasing and concave function with respect to the buffer size; when the buffer size over a threshold, the overflow probability is a monotonic decreas-ing and convex function with respect to the buffer size; when the buffer size is beyond a value, the average queue length is an approximately increasing linear function with respect to the buffer size; the average queueing delay is a linearly increasing function with respect to the buffer size. With the consideration of finite buffer, lossy link and signaling overhead, the transmission strategy without scheduling outperforms the existing best two-way relay protocols, and the throughput gain reaches nearly 20%. The derived Lemmas demonstrate that, increasing the buffer size is able to improve the system throughput efficiently, and the throughput gain is about 9.5%.b) A Grouped Markov model was proposed to capture the two-way relay with the consideration of scheduling. Currently, there was no existing analytic model to characterize the two-way transmission protocol with the scheduling consideration. To this end, a transmission protocol based on the scheduling and sending window was presented. The events associated to the relay and user transmission reside on different Markov layers, which were then projected into a new 2-D Markov model. This two layer Markov model is able to distinguish the ambiguous states caused by the transmission of the relay and the users. With the Gram-Schmidt algorithm factorization, we also present a simplified method to solve the state steady probability.c) An effort to optimize transmission method was made under the assumption of fixed rate transmission and fading channel. A condition for the optimization was constructed based on the queue theory, and was then converted into a KKT condition. By solving the optimization problem, the best transmission strategy was presented. The transmission strategy based on the KKT conditions is able to achieve 10% throughput gain compared with the stochastic network coding transmission. The closed form of the packet transmission successful probability was derived based on the outage probability, the ideal coding and modulation.d) A starving buffer method was proposed which can reduce the average queue length. The transmission protocol based on the KKT conditions requires a considerable transmission of the scheduling messages, which consume a lot of signalling resources, and the stochastic protocol causes a waste of the transmission resource. The starving buffer method is able to compromise the advantage and the disadvantage of the scheduling. Simulations demonstrate that the starving buffer protocol’s throughput lies between the KKT protocol and the stochastic protocol.With the consideration of the signalling overhead, starving buffer reaches a best throughput, and the throughput gain is nearly 14.3% compared to the KKT protocol.