A supervised learning approach to cognitive access point selection

In this paper we present a cognitive AP selection scheme based on a supervised learning approach. In our proposal the mobile station collects measurements regarding the past link conditions and throughput performance, and leverages on this data in order to learn how to predict the performance of the available APs in order to select the best one. The prediction capabilities in our scheme are achieved by employing a Multi-layer Feed-forward Neural Network (MFNN) to learn the correlation between the observed environmental conditions and the obtained performance. Our experimental performance evaluation carried out in a testbed using the IEEE 802.11 technology shows that our solution effectively outperforms legacy AP selection strategies in a variety of scenarios.