Link Quality Prediction for 802.11 MANETs in Urban Microcells

Abstract

In this paper we derive the optimal link quality predictor (LQPR) whose parameters are estimated from signal power and node speed samples. We propose a fast estimator for these parameters whose computational complexity is three orders lower than that of the optimal estimator with only a slight loss in accuracy thus enabling realtime execution. We show that using the most recent local mean of the signal as a predictor of future signal strength is also a very close approximation to the optimal predictor. This is the central result of this paper. It obviates the need for complex and/or computationally intensive link quality predictors for 802.11 in urban microcells and has the advantage of not requiring node speed information. The LQPRs are evaluated against the lower error bound. We show that the LQPR based on the most recent local mean of the signal predicts the packet reception probability for pedestrians in urban microcells on average with a mean absolute error of 13.47%, 16.54%, 18.21% and 19.38% for 1 s, 2 s, 3 s and 4 s into the future respectively. This LQP accuracy resembles closely the lower error bound with, for example, a difference of only 2.47% at 2 s into the future.

Wireless Body Area Sensor Networks: Survey of MAC and Routing Protocols for Patient Monitoring under IEEE 802.15.4 and IEEE 802.15.6

“…Additionally, low-power radios are very sensitive to noise, interference and multipath distortions. For the low-power WBASNs, link quality is considered a crucial metric for the selection of the next hop [ 152 , 153 , 154 ].…”