Article Preview
TopIntroduction
A WSN comprises of a large number of sensor nodes that are typically distributed over a wide area. The WSN also comprises of a node that acts as an information collector and is called the sink node. These sensor nodes usually consist of a sensor unit, a battery and a radio. In typical deployment scenarios, the sensor nodes may be deployed in remote, unattended, and harsh environment in large numbers. In some applications, the physical size of a sensor node is kept as small as possible for stealthy missions and cost cutting. In most cases, sensor nodes in WSN are equipped with batteries that cannot be recharged and have a limited lifetime. It may be difficult to recharge or replace their batteries due to their remote or hostile locations. This fact gives rise to major design goals for WSN, which is to reduce power consumption and to increase the life of sensor nodes as far as possible (Jones, Sivalingam, Agrawal, & Chen, 2001; Liu, Cao, Zheng, Gong, & Wang, 2008).
Data packets usually originate at the sensor nodes as a result of its sensing activities. Moreover, the sensor also behaves as a router to route the packets that arrive from other nodes. The data that are generated at the sensor nodes as well as the ones routed by the nodes must all reach a sink node in sensor network. The traffic in a wireless sensor network usually has a many-to-one pattern, where nodes that are located nearer to the sink bear a heavy burden of the traffic. As a result, it is seen that the nodes close to the sink would consume their power much faster, resulting in the eventual death of the node. This problem is known as the energy-hole problem (Jiang, Huang, Yang, & Leu, 2012). If an energy hole appears, then packets cannot be delivered to the sink anymore. The longevity of a WSN entirely depends on the life of sensor nodes that are closer to the sink of the WSN. As described in (Ye, Heidmann, & Estrin, 2004), there are four primary reasons of power consumption: collision, overhearing, control packet overhead, and idle listening.
Overhearing and idle listening use power by keeping the radio receiver in an ON state without actually getting any useful information. However, compared with the other three types of power consumption for a generic node, packet collisions lead to energy waste which is much more. Medium contention in a contention-based protocol is one of the main reasons of packet collisions among sensor nodes. Shih et al. (Shih, Cho, Lee, Calhoun, & Chandrakasan, 2004) have demonstrated that power consumption is very high in case of the radio transmitter switching from one mode to another. An approach to alleviate total average times of both medium contention and mode switching of radio receiver is presented in (Jiang et al., 2012). The approach discussed here is based on N-policy M/M/1 queuing model.
The nodes located near the sink of the WSN use more energy than other nodes because of the larger forwarding burden. This would lead to a faster depletion of the battery and eventually to the failure of the WSN. From the angle of queuing theory, the larger forwarding burden implies that the average arrival rate of packets at the nodes closer to the sink is higher than that of nodes farther away from the sink. To alleviate this problem, an optimal N criterion is adopted for a queue-based scheme. A queue threshold, N, is specified which would imply that the node will wake up only after the node’s buffer has at least N packets and the node switches back to idle state when the buffer is empty. In this scheme, when the node buffer is filled with N packets, the sensor node triggers its transmitter and starts transmitting the packets at one go. The optimal value (N*) of N at which the sensor nodes use the least power is also calculated.
In this paper, the authors present the mathematical expressions for the queue-based approach and also data simulations with MATLAB tool are conducted on optimal queued values for mitigating power consumption. The authors have conducted simulation of the queue-based approach using MATLAB and the results are presented here that confirm the power saving achieved by the queue based scheme.