Research On Integrated Communications And Signal Processing For Distributed Networked Acoustic Sensing Of Ocean Current Field

Ocean current is one of the basic forms of sea water movement,which transports energy,heat,and mass in the ocean.Ocean currents have important impacts on global climate,fisheries,shipping,tidal power generation,offshore structures,ocean environment,and naval operations.The study of ocean dynamic processes(such as the temporal and spatial distribution of ocean current)not only helps to solve scientific problems in the physical oceanography,but also helps us to rationally exploit the ocean.Sound can travel long distances in the ocean,whose propagation is affected by the ocean,and in turn carries the information on the ocean.Distributed networked acoustic sensing(DNAS)of the ocean current field measures the two-dimensional horizontal ocean current velocity field within the network by using the current field estimation methods that are based on the reciprocal travel time differences between each pair of nodes.We study the integrated communication and signal processing for distributed network acoustic sensing of the ocean current field,focusing on the communications and networking and the current field estimation methods for both centralized and decentralized network architectures.First,we analyze the process and application requirements of the distributed acoustic measurement of ocean current field under the two network architectures.Then,we design the spatial reuse time-division multiple access(TDMA)based medium access control(MAC)protocols and analyze their performances under two network architectures with different implementation scenarios.A sea trial deployed a network architecture with a fusion center,showing the feasibility of the proposed protocols.For the distributed ocean current field estimation with a fusion center,a sequential estimation method based on time evolution model is proposed.We analyze the performance and applicability of the proposed method of different time evolution orders under various measurement and communication conditions.It is shown that the time evolution order of the optimal sequential estimation method is mainly related to the measurement error and sensing period;the sequential estimation method with a large time evolution order is suitable for heavily noisy measurement,while the one with a small time evolution order is suitable for the large sensing period.The simulation and sea trial data analysis show that the proposed method has better estimation performance than the non-sequential estimation method in most cases.For the distributed ocean current field estimation under the decentralized architecture,we propose a sequential estimation method based on the spatiotemporal evolution model,which can be implemented in a single time-scale or two time-scale manner.The estimation performance of the method under various measurement and communication conditions is analyzed.It is shown that the performance of the proposed method is better than that of the non-sequential estimation method with a fusion center and is close to that of the sequential estimation method with a fusion center based on the spatiotemporal evolution model;the proposed method is robust in presence of measurement noise and random packet loss during the probe signal broadcasting while also having a certain tolerance for the random communication packet loss.We also propose a new consensus matrix to relieve the performance degradation of the decentralized method due to communication packet loss.We propose a DNAS system framework for ocean current field based on square division(SD)of the topology,addressing the problem of triangle division(TD)based signal processing method that the performance depends heavily on the spatial variation of current velocity and reducing the system power consumption.The simulation shows that the SD method has lower node-averaged power consumption,is less sensitive to the spatial variation of current velocity,and has better estimation performance than the TD method under various measurement and communication conditions.The sea trial data analysis show that the SD method is feasible in the ocean environment.We explores the applicability of DNAS of ocean current field in the polar environment.Using the Bellhop model and the reflection coefficients computed by under-ice acoustic propagation models,the influence of the ocean current on the sound propagation is simulated with the help of effective sound speed profile.Path-integrated current velocity under ice is obtained by inverting the reciprocal acoustic travel time difference.In addition,we describe the implementation of the DNAS system in the polar environment from the aspects of network topology and signal processing method,node spacing and(down)sampling frequency requirements,node synchronization,etc.,to clarify the application of DNAS for ocean current field measurement in the Arctic ice-covered areas.We study the communication and networking process and signal processing process of distributed acoustic measurement of ocean current field with a focus on the influence of communication conditions and network topology on the performance of ocean current field estimation.This thesis provides a theoretical basis and technical support to the integrated communication and signal processing for DNAS of the ocean current field.