An optimized cluster-based WSN design for latency-critical applications

Markov-based analytical modeling has been used extensively to model the operation of the MAC protocol of the IEEE 802.15.4 standard under diverse assumptions. These models reveal the parameters that control the behavior of each node in the network such as end-to-end latency, reliability and power consumption. Focusing on improving a certain metric is highly dependent on the type of application the Wireless Sensor Network (WSN) is designed to support. Although reducing power consumption is a primary design factor in WSNs, the emergence of delay and reliability critical applications such as Smart Grid, healthcare, and the Intelligent Transportation System (ITS) calls for more stringent latency and reliability considerations. In this paper, we develop an optimization model for clustered WSNs to minimize the end-to-end delay and power consumption while maintaining certain levels of reliability. We formulate our optimization problem based on mathematical expressions derived from a Markov-based model. Our results indicate that the lower bounds to expect on latency and power consumption, under constraints on reliability and WSN cluster size, are within the limits required by latency-critical applications.