Channel sensing strategy for channel load estimation

Abstract

The emergence of multi-channel wireless networks and cognitive radio networks has rendered dynamic channel selection an important task, and many existing channel selection schemes factor in the amount of wireless activities that take place in each channel, or channel load, to achieve load balance and maximize the utilization of wireless resources. In such environments, a monitoring node must sense the channel to estimate the channel load, yet when the node is equipped with a single radio interface, in which case the lone interface must be used for both channel sensing and data communication, there can be only a fixed amount of time allotted to channel sensing. In this paper, we show that a careful scheduling of channel sensing is needed to improve the accuracy of channel load estimation, and based on our findings, we devise a sensing strategy that minimizes the estimation error. Evaluation shows that our scheme can reduce the relative estimation error by as much as 40% in a heavily loaded channel environment.