Murali Tummala

Connectivity in Sensor Networks

We analyze the probability that an arbitrarily selected sensor node in a sensor network is connected to a specified number, m, of other sensor nodes; i.e., that a communication path exists between a selected node and at least m other nodes. Specifically, we consider the following problem: If given a collection of sensor nodes, each with a communication radius r, that are to be uniformly distributed over an area of size A, how many nodes must be deployed to assure that any node is connected to at least m other nodes with a specified probability? Alternatively, if we deploy a fixed number of nodes over a region, what is the probability that a selected node will have m neighbors? Such questions are particularly pertinent for unattended sensor networks that perform cooperative operations requiring the participation of at least m nodes

Optimal Preamble Sampling for Low Power Wireless Networks

Preamble sampling (or low-power listening) is a low-power technique used in wireless networks to allow nodes to check for waiting traffic by periodically waking up from a low-power sleep state. An increasing number of proposals exist in literature to provide static and adaptive solutions to implement preamble sampling. This paper tackles the important question of what is the optimal preamble sampling period in a multi-node wireless environment. Expressions are derived for the optimal sampling period for both contention-based and non-contention-based medium access schemes and the accompanying energy consumption analysis is provided. This work clearly identifies the optimal sampling period to achieve minimum energy consumption on a per node basis as a function of the size of the neighborhood, the packet arrival rate, the parameters of the preamble sampling scheme and accompanying medium access scheme, and the device parameters of the wireless node. Simulation results are included to support the findings.