Distributed Signal Processing And Information Spreading In Wireless Networks

With the development of MEMS technology and more highly integrated single-chip solu-tions, the wireless terminals have become much smaller and diversified. A wireless terminal can be a hand-held cellphone, tablet, laptop, and even a self-controlled mobile sensor node or vehicular node. Wireless access has become ubiquitously available with the popularization of Machine-to-Machine (M2M) networks, making things "inter-connected" even scattered all over the world. In summary, the key features of next-generation wireless networks are con-cluded as "portable and mobilized", with the capability of accessing the Internet anytime and anywhere. After some generalization, we can easily discover that, unlike the conventional cen-tralized network, the distributed network or its distributed feature is gradually replacing some of the traditional network features and playing a more important role. However, most existing communication architectures, including their embedded protocols and algorithms, are designed for centralized networks by default. It is impossible to directly upgrade the incumbent network-s, in order to accommodate the incoming large-scale and fully distributed networks, without a complete re-design or re-invention of the topology control, signal processing and information spreading methods. Meanwhile, the theoretical researches on large-scale and distributed net-works in existing literatures can barely meet the demands. Therefore, this thesis mainly focuses on the future distributed networks, and tries to explore some fundamental technologies. Specif-ically, we have investigated distributed topology control scheme, networked signal processing algorithms, distributed resource allocation methods, as well as distributed information spreading strategies. The main contributions are as follows:We develop distributed topology maintenance and distributed signal processing schemes. On the one hand, we provide a solution to the energy efficient spectrum-aware CRSN cluster-ing problem. Specifically, we first design clustered structure and model network-wide energy consumption and determine the optimal number of clusters. Then, we employ the ideas from constrained clustering and propose both centralized spectrum-aware clustering (CSAC) algorith-m and distributed spectrum-aware clustering (DSAC) protocols. Through extensive simulations, we demonstrate that DSAC can effectively form clusters under dynamic spectrum-aware con-straint. Moreover, it exhibits preferable scalability and stability with its low complexity and quick convergence under dynamic spectrum variation. On the other hand, a novel distributed compressed wideband sensing scheme for Cognitive Radio Sensor Networks (CRSN) is pro-posed. Taking advantage of the distributive nature of CRSN. the proposed scheme deploys only one single narrowband sampler with ultra-low sampling rate at each nodes to accomplish the wideband spectrum sensing. First, the practical structure of the compressed sampler at each n-ode is described in detail. Second, we show how the Fusion Center (FC) exploits the sampled signals with their spectrum randomly-aliased to detect the global wideband spectrum activity. Finally, the proposed scheme is validated through extensive simulations, which shows that it is particularly suitable for CRSN.We study the cognitive multiple access channel in CRSN. On the one hand, a novel concept of Joint Source and Channel Sensing (JSCS) is introduced. Every sensor node has two basic tasks:application-oriented source sensing and ambient-oriented channel sensing. With in-depth exploration, we find that these two tasks are actually interrelated when taking into account the energy constraints. The main focus is to minimize the total power consumed by these two tasks while bounding the distortion of the application-specific source information. Firstly, we present a specific slotted sensing and transmission scheme, and establish the multi-task power consump-tion model. Secondly, we jointly analyze the interplay between these two sensing tasks, and then propose a proper sensing and power allocation scheme to minimize the total power consumption. Finally, simulation results are given to validate the proposed scheme. On the other hand, we consider the fundamental problem of rate regions achievable for multiple secondary users (SUs) which send their information to a common receiver over such a white space channel. In partic-ular, the PU activities are treated as on/off side information, which can be obtained causally or non-causally by the SUs. The system is then modeled as a multi-switch channel and its achiev-able rate regions are characterized in some scenarios. Explicit forms of outer and inner bounds of the rate regions are derived by assuming additional side information, and they are shown to be tight in some special cases. An optimal rate and power allocation scheme that maximizes the sum rate is also proposed. The numerical results reveal the impacts of side information, chan-nel correlation and PU activity on the achievable rates, and also verify the effectiveness of our rate and power allocation scheme. Our work may shed some light on the fundamental limit and design tradeoffs in practical cognitive radio systems.We investigate the phenomenon of information spreading, rather than throughput and en-ergy consumption, in distributed and mobile networks. Specifically, we analyze the effect of mobility on information spreading in geometric networks through natural random walks. Our contributions are twofold. Firstly, we propose a new performance metric, mobile conductance, which allows us to separate the details of mobility models from the study of mobile spread-ing time. Secondly, we explore this metric for several popular mobility models, including the fully random mobility, the partially random mobility, the velocity constrained mobility, the one- dimensional area constrained mobility and the two-dimensional area constrained mobility, and offer insights on the corresponding results. We develop a large scale network simulation platform which is specifically used to verify our analysis. Note that our results and method are general, in order to adapt to various mobile networks, which make this work more significant and of more practical value.