Noise-Aware Energy-Efficient Sensor Binding

Most wireless sensor network (WSN) architectures involve clusters of sensor nodes reporting their sensed data to base stations or sink nodes, and end applications access sensor data from the sink nodes through service-oriented paradigms. One of the key challenges involved in the design of WSN applications through such architectures is the need to conserve the energy of batteries used to power sensor nodes. This is because sensor-to-sink communication takes up the bulk of the available battery power on a sensor node. In this paper, we present a novel strategy for associating a sensor node to one of many possible sink nodes. The problem is approached from an "energy efficiency" point of view in a noisy environment. We formulate cost metrics that can be utilized to evaluate what impact both signal to noise ratio (SNR) and interference level have on packet loss when a particular sensor node is associated with a particular sink node. Using these cost metrics, a sink node advices a sensor node wishing to associate with it on what costs would be incurred, and the sensor nodes chooses to bind with the sink node that has the least cost. We then undertake a simulation study to demonstrate the efficacy of our cost metrics towards energy efficiency and show that an implicit added benefit is the distribution of the total load on all the available sink nodes resulting in "load balancing".