Performance analysis of adaptive clustering in Hybrid Technology Networking

North America's railroad industry and the Federal Railroad Administration continuously strive to improve operational safety and security of their operations. In order to design a reliable and efficient real-time monitoring and control network for freight trains and railcars, we had previously proposed a novel approach called Hybrid Technology Networking (HTN). HTN deploys sensor nodes in clusters, with intra-cluster communication provided by ZigBee and inter-cluster communication accomplished via WiFi through dynamically assigned node roles called gateways. ZigBee provides HTN with low-power operations and WiFi enables long-range communication for high performance. Due to a complex working environment the network topology changes frequently, which often leads to poor resource utilization, reduced network lifetime, fluctuations in channel conditions, high latency in end-to-end data delivery and variable loading at the gateways. To resolve this problem, HTN provides for cluster adaptation. In this paper, we describe algorithms to merge and split HTN clusters for making them adaptive to changing operating scenarios. We also explore how the process of making clusters adaptive will affect latency and reliability in the network.