ABSTRACT
Cognitive radios are a disruptive technology based on the fundamental software-defined mechanism that enables wireless devices to sense their environment, and dynamically adapt the transmission method on the currently available spectrum. When such wireless devices are embedded in vehicles, the single-node challenges of dynamic radio resource allocation are exacerbated by the dynamic nature of the network. A vehicular network may have in fact variable contact rates, variable hop-to-hop and end-to-end delays, variable traffic density, and even variable goals, given by each service or application. In such complex heterogeneous and dynamic scenarios, a “one-size fits all” spectrum allocation solution possibly cannot exist.
To this end, in this chapter we propose two policy-based solutions that allow cognitive vehicular networks to distributively adapt locally to the available spectrum and globally to a service goal or objective by merely instantiating a few policies. In particular, in the first part of the chapter we model the spectrum allocation as a global utility maximization problem. We then show a policy-based technique that leverages decomposition theory to tune the behavior of the multi-dimensional spectrum allocation solution. In the second part of the chapter we focus on another policy-based technique. Leveraging recent results on the consensus literature we propose an approach that enables cognitive radio-equipped vehicles to allocate the spectrum in a distributed fashion, providing guarantees on both convergence time and performance with respect to a Pareto optimal spectrum allocation.
