An empirical study of link quality estimation techniques for disconnection detection in WBANs.

Sensor nodes in manyWireless Body Area Network (WBAN) architectures are supposed to deliver sensed data to a gate- way node on the body. To satisfy the data delivery require- ments, the network needs to adapt itself to the changes in connection status of the body nodes to the gateway. As a prerequisite, Link Quality Estimation (LQE) needs to be done to detect the connection status of the nodes. The qual- ity of links in WBANs is highly time-varying. The LQE technique should be agile to react fast to such link quality dynamics while avoiding frequent uctuations to reduce the network adaptation overhead. In this paper, we present an empirical study on using dierent LQE methods for detect- ing the connection status of body nodes to the gateway in WBANs. A set of experiments using 16 wireless motes de- ployed on a body are performed to log the behavior of the wireless links. We explore the trade-os made by each LQE method in terms of agility, stability, and reliability in detect- ing connection changes by analyzing the experimental data. Moreover, dierent LQE methods are used in an adaptive multi-hop WBAN mechanism, as a case study, and their impact on the Quality-of-Services (QoS) are investigated.