COOPERATIVE NETWORKING AND RELATED ISSUES: STABILITY, ENERGY HARVESTING, AND NEIGHBOR DISCOVERY

Abstract

This dissertation deals with various newly emerging topics in the context of cooperative networking. The first part is about the cognitive radio. To guarantee the performance of high priority users, it is important to know the activity of the high priority communication system but the knowledge is usually imperfect due to randomness in the observed signal. In such a context, the stability property of cognitive radio systems in the presence of sensing errors is studied. General guidelines on controlling the operating point of the sensing device over its receiver operating characteristics are also given. We then consider the hybrid of different modes of operation for cognitive radio systems with time-varying connectivity. The random connectivity gives additional chances that can be utilized by the low priority communication system.

The second part of this dissertation is about the random access. We are specifically interested in the scenario when the nodes are harvesting energy from the environment. For such a system, we accurately assess the effect of limited, but renewable, energy availability on the stability region. The effect of finite capacity batteries is also studied. We next consider the exploitation of diversity amongst users under random access framework. That is, each user adapts its transmission probability based on the local channel state information in a decentralized manner. The impact of imperfect channel state information on the stability region is investigated. Furthermore, it is compared to the class of stationary scheduling policies that make centralized decisions based on the channel state feedback.

The backpressure policy for cross-layer control of wireless multi-hop networks is known to be throughput-optimal for i.i.d. arrivals. The third part of this dissertation is about the backpressure-based control for networks with time-correlated arrivals that may exhibit long-range dependency. It is shown that the original backpressure policy is still throughput-optimal but with increased average network delay. The case when the arrival rate vector is possibly outside the stability region is also studied by augmenting the backpressure policy with the flow control mechanism.

Lastly, the problem of neighbor discovery in a wireless sensor network is dealt. We first introduce the realistic effect of physical layer considerations in the evaluation of the performance of logical discovery algorithms by incorporating physical layer parameters. Secondly, given the lack of knowledge of the number of neighbors along with the lack of knowledge of the individual signal parameters, we adopt the viewpoint of random set theory to the problem of detecting the transmitting neighbors. Random set theory is a generalization of standard probability theory by assigning sets, rather than values, to random outcomes and it has been applied to multi-user detection problem when the set of transmitters are unknown and dynamically changing.