Revisiting Link Quality Metrics for Wireless Sensor Networks

Packet Reception Ratio (PRR), Received Signal Strength Indicator (RSSI) and Link Quality Indicator (LQI) are common metrics for link quality estimation. However, utilization and statistical methods of these metrics are different, so it fails to describe their link quality estimation capabilities systematically and deeply. Some works even came to contradicting conclusions. In this paper, these three metrics are comprehensively evaluated through collecting and analyzing large amounts of experimental data. It is shown that average PRR could be used to distinguish good links from bad links. Standard deviation of PRR could be used to identify moderate links. So, good links, moderate links and bad links are able to be distinguished more effectively and accurately by combining average value and standard deviation of PRR. Both average RSSI and LQI could be used to identify good links. However, they could not distinguish moderate and bad links. Standard deviation of RSSI doesn't have link quality estimation capability, and so does the standard deviation of LQI within fixed time windows. Unlike them, standard deviation of LQI with fixed number of received packets could be used to identify good links, but still not moderate and bad links.