An Intelligent Packet Drop Mechanism in Wireless Body Sensor Network for Multiple Class Services Based on Congestion Control

Abstract

Considering the impact of congestion on wireless body sensor network (WBSN), an intelligent packet drop mechanism for multi-class traffic based on exponential random early detection is presented to mitigate the network congestion. A learning automaton is set in mote for “learning” intelligently from outside network environments. Obtained the learning result, an exponential random early detection algorithm is used in interactive node to control congestion by dropping packet. Meanwhile, based on the characteristics of WBSN, we also subdivide the packet dropping probability by the different priority traffic. Eventually, the packet dropping probability of traffic under different network environments, queue lengths and priorities is obtained. In consideration of a large amount of redundant data in WBSN, the proposed mechanism is able to actively dropping some low priority data, which can not only mitigate network congestion but also ensure the transmission of important vital signals.