BASS

Abstract

Whereas energy efficiency in wireless sensor network is of critical importance, idle listening has been recognized as a main source of wasted power. Many studies have proposed various approaches to scheduling the active and sleeping periods of sensor nodes, as to reduce the power consumption of idle listening. While noticing the fact that there is no universally accepted approach, one which can meet the diversity of different application, in this paper, we propose a scheduling scheme for active and sleeping periods that is based on the packet arrival pattern. More particularly, we propose an arrival model which is targeted at application characterized by bursty arrival. The bursty arrival times are assumed to be distributed exponentially with different rates for the packet arrival intra-burst and inter-burst. Based on this packet arrival model, we introduce a Bursty Arrival Dependent Sleeping Scheduling (BASS) scheme, in which each node dynamically and independently adjusts its wakeup rate. Through analysis and simulation, we evaluate the impact of the proposed scheme on the duty cycle and on the delay of the MAC layer. We show that as the bursty arrival rate decreases, the ON/OFF duty cycle decreases linearly and the MAC-layer delay is minimally affected. Our results suggest that the BASS scheme provides a superior solution for sensor network with bursty arrival. Comparing BASS scheme with S-MAC, results demonstrate 45%—70% gains in the BASS case in term of energy efficiency, without degrading performance. BASS scheme also obtains much better performance (e.g. 10 times better in the range of parameters in this paper) with the same amount of power consumption.