Optimizing Channel Access for Event-Driven Wireless Sensor Networks: Analysis and Enhancements
Abstract
We study the problem of medium access control in domain of event-driven wireless sensor networks (WSNs). In this kind of WSN, sensor nodes send data to sink node only when an event occurs in the monitoring area. The nodes in this kind of WSNs encounter correlated traffic as a subset of nodes start sending data by sensing a common event simultaneously. We wish to rethink of medium access control (MAC) for this type of traffic characteristics. For WSNs, many existing MAC protocols utilize the basic CSMA/CA strategy such as IEEE 802.11 Binary Exponential Backoff (BEB) algorithm to handle the collisions among packets when more than one node need to access the channel. We show that this BEB algorithm does not work well without incurring access delay or performance degradation due to increased number of collisions and retransmissions when nodes encounter correlated traffic. Based on above observations in mind, We present a Adaptive Random Backoff (ARB) algorithm that is capable of mitigating the impact of correlated traffic and capable of minimizing the chance of collisions. ARB is based on minor modifications of BEB. We show using numerical analysis that our proposals improve the channel access in terms of latency, throughput, and frame dropping probability as compared with IEEE 802.11 DCF. Simulations using NS-2 network simulator are conducted to validate the analytical results.