Adaptive interface reconfiguration of mesh WPANs in IOT environment

Abstract

The Internet of Things is emerging as global machine to machine and machine to people network. Mesh WPAN due to its self-sustaining nature and ubiquitous coverage makes it suitable for IOT remote devices (e.g. sensors) connectivity to IP based networks. In mesh WPANs, the data is propagated to the IP network in ad-hoc fashion through hop-to-hop communication. Denser deployments of remote devices are expected in Internet of Things that produces a huge amount of data. Traditional single interface sensor nodes restrict the whole network on a single channel and hence the full potential of available multiple channels is not achieved. The goal of this paper is to achieve high performance under the uneven distribution of traffic load, while also taking into consideration energy consumption. In this paper, second generation multi-interface sensor nodes are exploited for developing a mesh WPAN. Bayesian estimation is proposed to infer link capacity and accordingly a reconfiguration algorithm is presented for multi-interface sensor nodes. The proposed approach is evaluated through traces on multiple flows, collected from simulation. The performance evaluation demonstrates that link critical reconfigurations significantly improve network capacity in diverse traffic conditions.