Rate Measurement Over Short Time Scales in Stationary Cellular Receivers

An increasing number of cellular congestion control algorithms (CCAs) are becoming reliant on measurements of the delivery rate observed at the receiver. Accordingly, early detection of changes in the receiver’s rate would improve the performance of such algorithms. In addition to CCAs, faster detection of rate can also benefit available throughput estimation tools that rely on rate measurements. The upper layers of a cellular receiver could achieve faster rate detection through rate measurements over short time intervals. However, for cellular receivers, upper-layer rate measurements over short time scales produce unreliable results due to the effect of underlying lower layer mechanisms such as scheduling and retransmissions. In this paper, we introduce a Kalman filter based rate estimation approach that reduces the variability observed in short time scale receiver rate measurements and allows faster rate change detection. We also integrate an adaptive mechanism to facilitate online estimations in a network with an unknown or changing characteristic.