Abstract
Several key exchange methods for wireless channels have been presented in the literature. They are collectively named as PHY-layer security techniques and are usually based on the channel's fading characteristics and the reciprocity principle. In this paper, we present a PHY-layer security algorithm which includes channel sampling and thresholding and error reconciliation based on neural networks. The presented algorithm's performance and offers a security and these are studied with the help of simulations and interesting results are drawn about its overall utility.