Joint learning-detection framework: an empirical analysis

Abstract

Within a context of Cognitive Radio, we explore the possibility to detect the presence of users on a given band with no a priori knowledge on the environment. This paper ventures the analysis of a joint learning-detection framework where the detector improves its detecting abilities online relying on past gathered information on the probed frequency band. Simulations conducted on real data to support the suggested approach show interesting results. These results are then discussed and further possible improvements are proposed.