Opportunistic spectrum access in wireless communications with network coding

We consider opportunistic spectrum access on a channel with stochastic traffic carried by network coding. We show how a secondary user (SU) can leverage the structure induced by block-based network coding on a primary user's (PU) channel to maximize the throughput. Network coding enhances transmission efficiency and when applied on a PU channel, it can extend spectrum availability for the SU. We study the additional gain of spectrum predictability from network coding and show that the SU can more reliably detect the idle spectrum under sensing errors if the PU channel carries network-coded transmissions even when the channel utilization remains the same. The SU maximizes the throughput by first learning the PU spectrum parameters with the Baum-Welch algorithm and then tracking the spectrum holes with the Partially Observable Markov Decision Process (POMDP) algorithm. We show that network coding renders the spectrum more predictable, which leads to a higher SU throughput.