EAGER: NeTS: Under-Ice Mobile Networking: Exploratory Study of Network Cognition and Mobility Control

具有声学通信能力的自主水下航行器（AUV）是冰下勘探的首选平台。与通常研究的开放水域环境不同，冰下环境中的声速随水深的增加呈现出增大的趋势，这使得声传播受到冰盖的遮蔽和多次反射。在冰下声学通信系统中，必须审慎地考虑这种声学环境特性，否则可能导致如在现场实验中观察到的严重通信断开。该项目的重点是冰下AUV网络迁移作为一个群体的水采样在一个未知的冰覆盖区域，并开发算法的AUV学习冰下的声学环境和适应AUV的移动性的声学环境和水样场的特点，以实现最佳的冰下使命性能，同时保持所需的声学连接。该项目将扩大自动驾驶汽车在冰下探索的前沿。鉴于冰雪覆盖地区在经济增长和科学研究等现代社会的许多基本因素中发挥着至关重要的作用，该项目将产生重大的社会经济影响。此外，该项目还将资助两名博士。论文，并涉及初级研究人员在算法开发和现场实验。该项目将在两个相互关联的领域进行创新：冰下声学环境和网络认知，以及自适应AUV移动控制。具体而言，将开发一种递归算法来估计与声传播有关的环境参数，以及网络状态（包括AUV位置和速度），利用AUV网络内数据包传输期间获得的声学测量。所估计的参数将表征冰下声场的AUV的移动性控制。此外，一个自适应算法将被设计来调整的AUV的移动性的声场和水样场，以最大限度地减少样本场估计误差，同时确保AUV之间的期望的声学连通性的目标。所开发的算法将通过仿真和离线实验数据处理进行评估。本项目将在当地湖泊约10个月的冰盖期内，在广泛的几何和环境条件下进行广泛的冰下实验。该项目将开发和展示冰下AUV移动的网络的基本和交叉技术，支持环境认知，统计信号处理和无线移动的网络的协同作用。

Autonomous underwater vehicles (AUVs) with acoustic communication capabilities are the platform of choice for under-ice exploration. Different from commonly studied open-water environment, the sound speed in the under-ice environment exhibits an increasing trend with water depth, which renders sound propagation shadowing and multiple reflections by the ice cover. Such acoustic environment characteristics have to be judiciously accounted in under-ice acoustic communication systems, which otherwise could lead to severe communication disconnection as observed in field experiments. This project focuses on an under-ice AUV network that migrates as a swarm for water sampling in an unknown ice-covered region, and develops algorithms for AUVs to learn the under-ice acoustic environment and adapt AUV mobility to the characteristics of the acoustic environment and the water sample field to achieve optimal under-ice mission performance while maintaining desired acoustic connectivity. This project will expand the frontier of under-ice exploration by autonomous vehicles. Given the vital role of ice-covered regions in many underpinning factors of modern society, such as economic growth and scientific research, this project will yield significant socio-economic impacts. In addition, the project will support two Ph.D. dissertations, and involve junior researchers in both algorithm development and field experiments. This project will innovate over two interrelated domains: under-ice acoustic environment and network cognition, and adaptive AUV mobility control. Specifically, a recursive algorithm will be developed to estimate the environment parameters pertaining to acoustic propagation, as well as the network state (including AUV positions and velocities), leveraging the acoustic measurements obtained during packet transmissions within the AUV network. The estimated parameters will characterize under-ice acoustic field for AUV mobility control. Moreover, an adaptive algorithm will be designed to adjust the mobility of AUVs to the acoustic field and the water sample field, with a goal of minimizing the sample field estimation error while ensuring desired acoustic connectivity among the AUVs. The developed algorithms will be evaluated via simulations and offline experiment data processing. Within an about 10-month ice-cover period of local lakes in this project, extensive under-ice experiments will be conducted under a wide range of geometric and environment conditions. This project will develop and showcase fundamental and crosscutting techniques for under-ice AUV mobile networking, underlying the synergy of environment cognition, statistical signal processing, and wireless mobile networking.