Abstract
Nowadays there is a plethora of wireless handsets in the market such as smartphones, tablets, laptops and wearable devices, that together with the the future emerging scenarios on vehicle to vehicle communications and smart city infrastructure will populate urban environments with a broad diversity of multi-standard wireless devices. This increase in the density and diversity of mobile devices have been the driver for collaborative protocols that can deliver effective communications. Cooperation is a technology that has the potential to provide energy efficient and scalable communications, where nodes play an important role to coordinate local traffic and act as gateways to the core network. Despite the diverse power requirements of multi-standard wireless interfaces and the different channel characteristics, support for energy efficient communications where relay nodes can be selected with lower energy requirements or with higher order modulation opportunities, is still expected. In this framework, clustering is a widely accepted technique that allows nodes create and join virtual cooperative groups, and to select a clusterhead that can provide a high speed and energy efficient backhaul link to the mobile network. The vast majority of existing clustering techniques assume that the collection of nodes that form a cluster are either static or have very low relative velocity. However, in practice nodes or devices are constantly on the move providing the impetus for mobility aware clustering techniques that elect a subset of nodes with a common mobility pattern. In this context, mobility-aware clustering, based on geolocation, is an active field of research due to the increasing interest of vehicular communication technology. However, clustering has a wide range of applications where GPS information is not always available. This requires a new design of clustering algorithms that do not depend on GPS coordinates. This challenge has fostered a new vision of clustering based on cognition, where nodes form mobile clusters that can adapt on-demand to the scenario characteristics. This thesis investigates cluster formation exploiting the notion of wisdom of crowds, where the nodes are aware of the surrounding mobility patterns and can adapt the cluster formation strategy to suit the current mobility trends. Moreover, this thesis also caters for a novel analytical model for cluster lifetime that is used to validate our simulation results. Another dimension to the clustering problem is how to exploit available spectral opportunities for cluster formation in a secure manner. Cognitive radio, and more concretely cooperative spectrum sensing is evaluated in this thesis as a solution for data channel assignment in mobile clusters. In this scenario, we focus on the security concerns of cooperative spectrum sensing. Namely, we address spectrum sensing data falsification and incumbent emulation attacks, and propose an energy efficient security mechanisms based on lightweight cryptography to address these threats.