A comparative study on the performance of wireless sensor networks

To increase the lifetime of the sensor networks, a recognized method is to switch off/on some sensor nodes between "sleep" and "activity" mode in order to save the energy. The performance metrics of an individual node can be calculated by using its DTMC model, the theoretical results include the average number of data units generated in a time slot lambda(E), the sensor throughput T, the average buffer occupancy B. In our research, we have implemented a wireless sensor network on NS2 in which, the sensors nodes can be either in sleep or activity modes. Our simulations have produced the experiment results of performance metrics. By comparing the experiment results with the theoretical results, we demonstrate the DTMC model is able to accurately describe the behavior dynamics of a sensor node.