(Ne)tworking (W)ireless (S)ensors Lab

Home Up Feedback Contents Search


Abstract

 

 

 

 

 



Energy Efficient Communication in Stationary Wireless Sensor Networks

(Ph.D. Thesis, Ayad Salhieh)


Wireless sensor networks have become possible because of the on-going improvements in sensor technology and VLSI. An important issue in smart sensor networks is achieving efficient operation because of the limited available power. Energy conservation is a critical issue in wireless sensor networks for nodes and network life, as the nodes are powered by batteries only. In order to maximize the lifetime of the nodes and the network, the traffic should be routed such that the energy consumption is balanced among the nodes in proportion to their energy reserves, instead of routing to minimize the absolute consumed power for each message.

In this dissertation we focus on three major points toward an efficient communication for stationary sensor networks. First, we study the relationship between the choice of topology and the power dissipated in the network. Second, we evaluate the effect of power metrics on the choice of routing and the power dissipated in the network. Finally,we determine the effect of using local information instead of using global information on extending the lifetime of the network.

Routing packets within a large-scale wireless sensor network without storage overhead and routing table updates is a challenging problem. With a large number of sensors the overhead plays a significant role in the scalability of the routing protocol. In order to avoid this communication overhead, sensor network routing demands new and efficient methods for routing packets. In order to remove or reduce this overhead, the routing protocol needs some way of implicitly, rather than explicitly, defining paths.

We introduce first the idea of directional routing, which requires only that each sensor know its location within the network relative to the sending node and the destination. This allows the use of simple directional routing based on local information only. For dense networks, sensors near a trajectory from the source to the destination can be found. However, sensors are energy-constrained devices, so selecting paths within this network could benefit from an energy-aware routing process.

Secondly, the problem of selecting paths leads us to examine the relationship between power usage and the number of neighbors in a wireless sensor network. Selecting the number of neighbors controls the type of topology to be used. The question that we are seeking to answer is what is the best topology for a wireless network of sensors, assuming we can control the placement of these sensors and the sensor locations are fixed relative to each other. Because the number of neighbors differs with different topologies, one expects different topologies to have different power usage rates. Even our simulations of the contention-free case show that different topologies have different levels of power efficiency. The results show that the total power consumption is reduced for topologies with fewer neighbors even though topologies with more neighbors require fewer hops.

Third, a routing protocol for wireless sensor networks need to be power aware. In order to be power aware we evaluate a number of power-aware routing protocols based on local information only. The simulations show that basing the routing decision on the remaining power of neighboring nodes is not enough by itself. Instead, using the directional value and the sum of power remaining at the next neighbors gives the routing protocol a broader perspective about the condition of the network from a local point of view and enhances the routing decision process.

Finally, in  order to gain some understanding of the quality of these local metrics, we also compare the energy usage and path length of these local methods with respect to some routing techniques based on global information. This evaluation demonstrates that changing the routing metric can dramatically affect the performance of the sensor network. These results also shows trade-off between extending the lifetime of the sensor network and reducing the average number of hops a message travels to the base station.

See full dissertation.

 

 

Home Up People Research Related Resources Sponsors

Send mail to manishk@cs.wayne.edu with questions or comments about this web site.
Last modified: February 10, 2002