An unscented Kalman filter based available bandwidth estimation algorithm for space bundle links

This paper presents an Unscented Kalman filter (UKF) based algorithm for estimating available bandwidth in a space Disruption-Tolerant Network (DTN) link. In the proposed algorithm, an UKF model is proposed for accurately tracking available bandwidth of bundle delivery over a space link. In particular, the proposed UKF will iteratively filter a serial of noised measurements on successive time intervals, which are obtained from a sequence of probe bundle pairs sent with a prior rate. Experimental results show that proposed algorithm can effectively estimate and track the available bandwidth over a time-varying space bundle link.