Cost-Effective, Cognitive Undersea Network for Timely and Reliable Near-Field Tsunami Warning

Abstract

The focus of this paper is on developing an early detection and warningsystem for near-field tsunami to mitigate its impact on communities atrisk. This a challenging task, given the stringent reliability andtimeliness requirements, the development of such an infrastructureentails. To address this challenge, we propose a hybrid infrastructure,which combines cheap but unreliable undersea sensors with expensive buthighly reliable fiber optic, to meet the stringent constraints of thiswarning system. The derivation of a low-cost tsunami detection andwarning infrastructure is cast as an optimization problem, and aheuristic approach is used to determine the minimum cost networkconfiguration that meets the targeted reliability and timelinessrequirements. To capture the intrinsic properties of the environment andmodel accurately the main characteristics of the sound wave propagationundersea, the proposed optimization framework incorporates the Bellhoppropagation model and accounts for significant environment factors,including noise, varying undersea sound speed and sea floor profile. Weapply our approach to a region which is prone to near-field tsunamithreats to derive a cost-effective under sea infrastructure fordetection and warning. For this case study, the results derived from theproposed framework show that a feasible infrastructure, which operateswith a carrier frequency of 12-KHz, can be deployed in calm, moderateand severe environments and meet the stringent reliability andtimeliness constraints, namely 20 minutes warning time and 99 % datacommunication reliability, required to mitigate the impact of anear-field tsunami. The proposed framework provides useful insights andguidelines toward the development of a realistic detection and warningsystem for near-field tsunami.