Analysis of Strategies for Mitigating the Sensor Netowrk Hot Spot Problem

In this work, we have shown the difficulty of mitigating the "hot spot" problem that arises in large scale wireless sensor networks when the data sink is a static node and when each node generates the same amount of data traffic. Although the most intelligent transmission power control policies are able to solve the problem by varying the transmission range for different nodes at different locations, it cannot significantly improve the network lifetime compared to the policy of uniform transmission powier. We argue that transmission power control alone is not enough to solve the "hot spot" problem and that either data sink rotation or data aggregation is necessary for the network to operate in an energy-efficient manner. In this work, we explore the former solution of optimizing data sink rotation and observe the benefits that can be derived from moving the data sink in the network. An alternative approach of deploying a small number of higher-powered aggregator nodes in different locations that can locally process sensor data is also investigated. However, since the movement of the data sink and the deployment of more aggregator nodes may be significantly more expensive than the deployment of an ordinary sensor node, there is a cost tradeoff involved in both approches. We explore how to balance the tradeoff to find a cost-efficient deployment plan.

On the Problem of Unbalanced Load Distribution in Wireless Sensor Networks

In multi-hop wireless sensor networks that are characterized by many-to-one traffic pattern, problems related to energy imbalance among sensors often appear. When each node has a fixed transmission range, the amount of traffic that sensor nodes are required to forward increases dramatically as the distance to the data sink becomes smaller. Thus, sensors closest to the data sink tend to die early, leaving areas of the network unmonitored and causing network partitions. Alternatively, if all sensors transmit directly to the data sink, the farthest nodes from the data sink will die much more quickly than those close to the sink. While it seems that a more intelligent transmission power control policy that requires nodes farther from the data sink to transmit over longer distances, such a policy can only have a limited effect, as energy balancing can be achieved only at the expense of gross energy inefficiencies. In this work, we investigate the transmission range distribution optimization problem and show where these inefficiencies exist when trying to maximize the lifetime of many-to-one wireless sensor networks.

For more information about this project, contact Mark Perillo or Zhao Cheng.