Two-way Relaying And Cross-layer Optimizations In Wireless Communication Systems

The two-way relay (TWR) protocols are efficient in providing appreciable throughput gains in wireless networks through the use of network coding to combine packets from multiple channels. The key determinant factor in driving the throughput improvement is the degree of simultaneity (DoS) achieved in the relay scheme. In this dissertation we propose a new two-way relay protocol named IC-TWR which combines network coding, spatial diversity, and interference cancellation (IC) techniques to arrive at high degree of simultaneity, and in the meanwhile to relax the requirement on channel state information (CSI) as compared with TWR schemes based on amplify-and-forward (AF). Numerical analysis shows that the proposed IC-TWR is uniformly advantageous over the traditional decode-and-forward (DF) scheme in terms of system throughput and end-to-end delay. The extension of IC-TWR protocol allows us the flexibility in selecting the candidate relay nodes in a given network topology.Assuming that the source nodes are backlogged and there are always sufficient data packets waiting to be transmitted, two-way relaying schemes have been extensively studied. However, this assumption is not always valid when queuing effects are taken into account and in practice the buffer size is always finite. In this dissertation, we analyze the joint effects of decode-and-forward two-way relaying and one-side finite-length queuing, which can be modeled as a two-dimensional finite state Markov chain (FSMC). We present a general analytical procedure and derive the closed-form expressions for system throughput and packet loss rate. Our theoretical analysis is verified by Monte Carlo simulations. Numerical results also illustrate how the buffer size influences the system performance in different scenarios. The framework may be helpful for system designers.Cross-layer design has been proposed to reduce the energy consumption in wireless sensor networks (WSN). The problem can be modeled as a hybrid non-linear optimization problem for integer constrains involving, which is a challenging task. As the number of nodes increases, the computational complexity grows exponentially. Constrained by computational complexity, previous work by other researchers mainly focused on no more than 4 nodes along a line (1-dimensional), so the advantages of cross-layer optimization cannot be fully explored and demonstrated. By solving the problem with mathematical models and numerical methods, we propose a detailed study on cross-layer optimization with up to 9 nodes in planar case (2-dimensional), which is more realistic. Numerical results show that optimal combination of multihop routing, link adaptation and variable-length TDMA schemes lead to significant energy savings in WSNs.