Efficient Delay-aware Data Collection in mostly-off Wireless Sensor Networks
Résumé
A challenging issue in data collection in duty-cycled wireless sensor networks (DC-WSNs) is the trade-off between lifetime and latency. Indeed, duty-cycling has long been a critical mechanism for energy conservation in WSNs where the sensor nodes switch between active and dormant states. Moreover, this mechanism involves strong sleep latencies leading to severe endto-end delay degradation. In this paper we propose a delayaware routing scheme to minimize the end-to-end delay in the data collection process already impacted by the significant sleep durations of the nodes. Most of traditional algorithms are based on restricting the propagation direction toward the sink from the propagation source in each hop and greedily searching the furthest neighbor each time. In our scheme we keep the same restriction on the propagation direction and we gradually adapt the transmission power allowing to each node to select all the potential forwarders regardless of their distance from the sink. First, we develop (FaST) a dynamic ForwArder Selection Technique where each node having packets to send must choose a forwarder whose next active time-slot is the closest. Besides, we combine power controlled unicast transmissions with FaST to better bound the communication delay while keeping energy consumption at acceptable values. Through extensive simulations, we examine the performance of our design Under different configurations and show how it outperforms traditional routing schemes in terms of end-to-end communication delay, load balancing and energy cost.