Efficient Topology Construction and Autonomous Networking for Underwater Acoustic Nodes

Abstract

Underwater acoustic networks operate in a dynamic and inhomogeneous environment where inhomogeneity is often apprehended by observing the Sound Speed Profile (SSP) of the water column. Moreover, underwater nodes often leverage directional transmission to increase the received signal strength while conserving a transmitter's energy. In addition, underwater nodes must establish and maintain a connected network topology in an ad-hoc manner while operating in a three-dimensional environment. Coupling the previous challenges with node drifts caused by unknown underwater currents complicates the process of identifying the location of neighbors and introduces challenges in maintaining underwater communication links. Fundamentally, the SSP causes acoustic signal refraction that may trigger a propagation path reversal in the vertical direction, thus complicating the ranging and localization process. Furthermore, utilizing directional transmission while nodes are drifting causes breaks in active communication links and thus nodes need to find new angles to reestablish these links. Moreover, selecting arbitrary transmission angles may lead to overlapping beams or result in leaving an underwater region uncovered. Finally, since SSP changes in the 3D environment control the propagation path and the spreading of the underwater beam, i.e. underwater losses, a node ought to have some means of estimating the SSP within its vicinity and in real time. This thesis addresses the above-mentioned challenges for establishing a robust underwater network topology in an inhomogeneous environment using directional transmissions. First, a transmission angle selection algorithm is proposed to enhance underwater spatial utilization while using directional transmission. Second, an effective distributive technique is developed to aid a node in orienting itself where the result ensures that all nodes across the network conform to the same axis. Third, to assist nodes in managing a connected network topology, a novel ranging process is proposed using a second order polynomial where the effect of medium inhomogeneity onto a propagating acoustic signal is assessed by observing the depth deviation among pairs and the corresponding change in the angles of transmission and reception. Fourth, an autonomous underwater sound speed estimation mechanism is introduced. Fundamentally, the trajectories of established acoustic links are utilized to infer the water column SSP where we leverage the fact that the slope of the propagation path has a direct relationship with the observed sound speed changes. Fifth, a novel ray classification mechanism is introduced where nodes identify the type of ray responsible for establishing a communication link with each neighbor and deduce an area where such ray types can sustain communication to the selected neighbor. Sixth, a computationally lightweight technique is presented to aid nodes in assessing the losses experienced by a link where the estimated losses are used to determine a maximum range based on the pursued ray type. Our proposed algorithms are distributed in nature and require no information about the underwater environment, but rather rely only on communication links formed during neighbor discovery. Finally, we validate our proposed algorithms using simulation and compare the results with competing schemes using actual field data.