An efficient technique for the reduction of packet loss in wireless sensor network

Abstract

This paper, presents an efficient technique for the reduction of packet losses in wireless sensor network (WSN), which occurs because of buffer overflow and subsequently congestion. A key design feature for wireless sensor network is to have less packet loss, as more packet loss degrades the Quality of service (QoS) of the transmitted information and retransmitting such, a packet can use up the energy of the nodes. Regardless of a number of efforts made to alleviate this problem, networks still experiences packet losses because of congestion and topology change. A congestion avoidance algorithm, which is based on an active queue management technique, was proposed in this work. The new technique, called a Refined Adaptive Random Early Detection (RARED) technique was used to control the maximum packet drop probability to be within 0.01 and 0.5, which is equivalently between 1% and 50%, thereby reducing the rate of packet loss and increasing the throughput. Simulation results obtained using Network simulator 2, showed that the proposed algorithm attained a reduction in packet loss of 5.7% against the existing Priority-based Fairness Rate Control (PFRC) in WSN.