Abstract
In cognitive radio networks which utilize unused spatial dimensions for secondary transmissions, estimating the number of streams provides essential information for the maximization of secondary throughput. By estimating the number of streams transmitted by the primary users, the secondary network can determine the number of spatial degrees of freedom that are available for opportunistic spectrum use. In this paper, we study the problem of streams number estimation in the context of cooperative cognitive radio network. We employ the well-known minimum description length (MDL) algorithm at secondary terminals and propose a decision fusion method which exploits the estimation bias of the MDL algorithm. The proposed method can provide improved estimation performance especially in the low signal-to-noise ratio region.
