Bound-Based Power Optimization for Multi-Hop Heterogeneous Wireless Industrial Networks Under Statistical Delay Constraints

TL;DR

This paper introduces delay-bound-based algorithms for power optimization in multi-hop heterogeneous wireless industrial networks, achieving significant power savings and improved network lifetime under statistical delay constraints.

Contribution

It develops novel power management algorithms using stochastic network calculus for multi-hop heterogeneous networks, addressing delay constraints and demonstrating substantial power savings.

Findings

Up to 95% power saving compared to fixed schemes.

Link heterogeneity reduces network lifetime without power management.

Proposed algorithms perform close to optimal in WirelessHART networks.

Abstract

The noticeably increased deployment of wireless networks for battery-limited industrial applications in recent years highlights the need for tractable performance analysis methodologies as well as efficient QoS-aware transmit power management schemes. In this work, we seek to combine several important aspects of such networks, i.e., multi-hop connectivity, channel heterogeneity and the queuing effect, in order to address these needs. We design delay-bound-based algorithms for transmit power minimization and network lifetime maximization of multi-hop heterogeneous wireless networks using our previously developed stochastic network calculus approach for performance analysis of a cascade of buffered wireless fading channels. Our analysis shows an overall transmit power saving of up to 95% compared to a fixed power allocation scheme when using a service model in terms of the Shannon capacity limit. For a more realistic set-up, we evaluate the performance of the suggested algorithm in a WirelessHART network, which is a widely used communication standard for process automation and other industrial applications. We find that link heterogeneity can significantly reduce network lifetime when no efficient power management is applied. Moreover, we show, using extensive simulation study, that the proposed bound-based power allocation performs reasonably well compared to the real optimum, especially in the case of WirelessHART networks.