Autonomous subsea platforms operating in the Arctic throughout the winter require a long-range acoustic navigation system to substitute for GPS while under ice and not able to surface. To allow the autonomous platforms to sample a large portion of the Beaufort, the navigation system must span several hundred kilometers, which in turn requires multiple sources with acoustic ranges of at least 100 km each. To enable vehicles such as gliders to navigate while performing their missions also requires an onboard acoustic receiver and accurate clock, plus processing to compute geolocations. Over the past four years an operational system that includes multiple fixed source moorings and compact, low-power receivers has been maintained in the deep area of the Beaufort Sea north of Alaska. The navigation system has been used to successfully allow gliders and profiling floats to operate throughout the winter, performing science missions and augmenting data collected on instrumented fixed moorings in the center of the Beaufort.

During a set of acoustic field experiments in the Fall River Harbor area of Massachusetts, data were collected to characterize underwater acoustic conditions for communications. A high-resolution bathymetric survey of the area was performed to create a gridded bathymetric product of the seafloor that was used for 3D acoustic propagation simulations in the given channel. The results were convolved with the actual transmitted waveforms to simulate received modem signals. In this paper, acoustic modem performance is compared with the modeled performance to examine the feasibility of the model for predicting acoustic performance based on understanding key system variables.

This letter presents a precise navigation technique for a scuba diver using an autonomous underwater vehicle (AUV) as a positioning aid. We develop and evaluate the state estimation algorithms and communication architecture for diver navigation based on subsurface human-AUV teaming with no requirement for ocean current data or exact diver speeds. By depending on acoustic communication and commercial AUV navigation capabilities, our method provides a unique capability for localizing a diver's position during an extended mission, without the requirement to maintain visual contact with the diver during the mission, or to deploy acoustic beacons. We utilize the Woods Hole Oceanographic Institution Micromodem 2 for range-only single-beacon navigation between two kayaks serving as proxies for the diver and REMUS 100 AUV. Range and odometry measurements are fused in a factor graph using incremental smoothing and mapping 2 (iSAM2) with appropriate motion and measurement models to provide real-time diver position estimates given unknown ocean currents. Field experiments demonstrate an average online endpoint error of 4.53 meters after 400-meter transits.

Autonomous systems, including gliders, floats, and propeller-driven AUVs, require navigation under Arctic ice to be able to geolocate during their missions. In addition, acoustic communications provides for unidirectional or bidirectional data and command flow with the autonomous systems. Over the past ten years, steady progress has been made on implementing and demonstrating these acoustic systems in the Beaufort Sea north of Alaska. The eventual goal is a multi-frequency capability that uses 35 Hz for very long range navigation, 900 Hz for long-range navigation and communications, plus 10 kHz for close-range use. The different frequencies offer different advantages, and each is used accordingly. At 35 Hz, navigation transmissions span the water-column and survive reflection from the underside of the ice, allowing pan-Arctic ranges. The 900 Hz navigation and communications signals persist in the shallow (150 m) duct in the Beaufort at ranges up to several hundred kilometers, which is useful for under-ice survey transits. Finally, at 10 kHz, compressed data can be off-loaded and new mission files transferred via buoys on the ice. Recent work in 2021 and 2022 has continued to develop these capabilities, and this talk summarizes results and describes future directions.

A software framework, “ros_acomms,” has been developed to enable transport of ROS messages and other data across low-throughput and high-latency underwater acoustic links. Messages are efficiently marshalled using user-provided configuration data, if available, or automatically via message introspection. A modular set of modem drivers, media-access-control engines, and message queues transport messages from one system to another via a modem. It supports message fragmentation, positive acknowledgment, and custody-transfer routing. It also includes an acoustic link simulator that uses a raytracing model to estimate link performance and latency. While it targets the WHOI Micromodem family of acoustic modems, the modular modem driver implementation has been leveraged to support low-throughput Iridium satellite links and other acoustic modems. It has been tested and used operationally at sea for remote redirection of autonomous underwater vehicles while providing operators with near real time vehicle telemetry and sensor data.

This paper presents improved algorithms for localization and navigation in which an autonomous underwater vehicle (AUV) supports a human diver. Our initial efforts validated state estimation algorithms and communication protocols for accurate diver navigation based on subsurface teaming with no ocean current data or exact diver speeds. By leveraging acoustic modem messaging and iterative ranging between the AUV and diver, this collaborative team maintains a loosely-coupled support structure that does not rely on close proximity or maintaining sight of a teammate. Range and odometry measurements comprise a factor graph structure that leverages the incremental smoothing and mapping 2 (iSAM2) algorithm for state estimation. However, this approach suffers from decreased accuracy in environments with heavy ocean currents. This requires an updated measurement strategy for ocean currents and new communication protocols to allow a diver to compensate for ocean currents. Extensive simulation results and comparisons to previous non-adaptive techniques show that these updates enable more efficient diver paths to a known target, decreased workload on the diver, and increased accuracy and robustness to ocean currents at the limits of human diver capability.

The under-ice acoustic transmission experiment of 2013, conducted under ice cover in the Fram Strait, was analyzed for bottom interactions for the purpose of developing a model of the seabed. Using the acoustic signals, as well as data from other sources, including cores, gravimetric, refraction, and seismic surveys, it was deduced that the seabed may be modeled as a thin surficial layer overlaid on a deeper sediment. The modeling was based on the Biot-Stoll model for acoustic propagation in porous sediments, aided by more recent developments that improve parameter estimation and depth dependence due to consolidation. At every stage, elastic and fluid approximations were explored to simplify the model and improve computational efficiency. It was found the surficial layer could be approximated as a fluid, but the deeper sediment required an elastic model. The full Biot-Stoll model, while instrumental in guiding the model construction, was not needed for the final computation. The model could be made to agree with the measurements by adjusting the surficial layer thickness.

We analyze data retrieved from an ocean floor pressure sensor continuously operated for 48 days in the Mentawai Strait during the Spring of 2016, as part of Project Hazard SEES. Initial processing through systematic spectrogram analysis has identified ten distant earthquakes recorded through the variation of pressure accompanying the passage of seismic waves on the bottom of the ocean. The analysis of the corresponding wavetrains allows the recovery of the standard magnitude $$M_s$$ of seven of the events (two more being intermediate depth, and the tenth antipodal) with a residual not exceeding 0.3 logarithmic units. We also show that the classical energy-to-moment ratio computation can be successfully adapted by defining a response function of the pressure sensor to teleseismic P waves. In addition, six local earthquakes, occurring at distances of 58–670 km from the sensor, but with moment magnitudes less than 5.7, were also recorded. We show that an estimate of the seismic energy radiated by these events can be obtained from a simple integration of the square of the pressure signal. Thus our results indicate that meaningful quantitative estimates of the source characteristics of both teleseismic and regional events can be obtained through robust methods based on single-station pressure recordings on the ocean floor.

The development of a synthetic baseline navigation technique that self-localizes an autonomous underwater vehicle (AUV) using intermittent acoustic communications signals received by a single transducer is described, along with field results from in-ocean tests. The method uses the phase measurement at the output of a second-order phase-locked loop to create fine-scale pseudorange estimates in addition to, or in the absence of, a one-way travel time measurement based on the arrival time of the acoustic data packet. These range measurements are incorporated by an adaptive particle filter. This technique allows the vehicle navigation system to take advantage of multiple phase-derived range measurements made over the duration of a communication packet. These measurements, when incorporated with an appropriate filter and vehicle kinematic model, improve vehicle navigation at no additional cost in navigation-specific acoustic transmissions. This approach was demonstrated and evaluated with data collected at-sea using a REMUS 100 AUV (Hydroid, Inc., Pocasset, MA).

A small experimental network of ocean sensors linked by acoustic communications was tested in March-April of 2017 in shore-fast ice near Thule Air Base in northern Greenland. The objective was to simultaneously test a real-time communications system for linking together under-ice sensors, and also to gather synoptic oceanographic and acoustic data to be used for improving under-ice propagation modeling. The acoustics portion of the experiment included tests at ranges from 3 to 35 km at a carrier frequency of 3.5 kHz at data rates from approximately 120-1200 bits per second. Initial results from the Micro-Modem real-time hardware used in the Thule test showed reliable links at 20-25 km at most data rates, and moderate connectivity from 25-35 km, depending on the specific path and data rate. The communications performance is governed by several factors including: the relative depth of source and receiver, the location within the fiords and proximity to glacial fronts with varying sound-speed profiles, pinniped (seal) vocalizations, pier-side machinery noise and occasional random impulsive noise events attributed to iceberg movement. The results show good link stability despite a spatially-variant sound speed field resulting from mixing of glacial and ocean waters.

Operating autonomous underwater vehicles at high latitudes is a challenge because ice cover prevents the use of GPS or data communications. As a result, our scientific observations are biased towards late spring, summer, and early autumn when ships can navigate and autonomous platforms can safely surface. To address this problem, we studied the feasibility of a basin-scale multipurpose acoustic network called the “Baffin Bay Acoustic Navigation and Communication System” (BBANC). BBANC would deploy broadband low frequency sources and receivers, offering one-way communication, acoustic positioning, and acoustic thermometry services. Passive acoustic listening elements would support the study of marine mammal communication and ambient noise from ships, ocean-based resource exploitation, and ice dynamics, as well as gate acoustic source operation in the presence of marine mammals. We describe the challenges and design parameters for such a system, as well as define additional acoustic and remote sensing measurements required to complete a system design. Drawing from a large database of Baffin Bay hydrography, we present simulations of under-ice sound speed conditions, ice properties derived from satellite remote sensing and upward looking sonar data, and modelled acoustic propagation paths in an ice-covered Baffin Bay. We also assess the feasibility of non-coherent and coherent communication.

A bottom-to-bottom acoustic communications network is feasible in areas of the ocean where the combination of the depth and near-surface sound speed support direct-path refracted rays which connect multiple sensor nodes located on or near the seafloor. In this paper the initial results of modeling and testing the physical layer of a communications system that will support a bottom-based sensor network in the Mentawai Basin west of Padang, Sumatra in Indonesia are reported based on an experiment done in 2015. The acoustic communications system utilizes a carrier frequency between 3 and 4 kHz, and achieves data rates from 60-700 bps, depending on the range. The combination of a bottom depth of 1750 m and a deep and seasonally persistent thermocline provide the necessary conditions for direct-path propagation via surface refracting rays. The results that are presented include signal-to-noise ratio (SNR), adaptive equalizer mean-square-error and sample multipath measurements from real-time acoustic modems (the WHOI Micro-Modem) at ranges of 20 and 27 km. Raw acoustic data was also recorded in parallel for post-processing and additional analysis. Refracted rays provided direct-path ranges to 33 km, so the results shown here can be extended through use of lower frequencies to increase the range.

A characteristic surface duct beneath the sea-ice in the Marginal Ice Zone causes acoustic waves to be trapped and continuously interact with the sea-ice. The reflectivity of the sea-ice depends on the thickness, the elastic properties, and its roughness. This work focuses on the influence of sea-ice roughness on long-range acoustic propagation, and on how well the arrival structure can be predicted by the full wave integration model OASES. In 2013, acoustic signals centered at 900 Hz were transmitted every hour for three days between ice-tethered buoys in a drifting network in the Fram Strait. The experiment was set up to study the signal stability in the surface channel below the sea-ice. Oceanographic profiles were collected during the experiment, while a statistical description of the rough sea-ice was established based on historical ice-draft measurements. This environmental description is used as input to the range independent version of OASES. The model simulations correspond fairly well with the observations, despite that a flat bathymetry is used and the sea-ice roughness cannot be fully approximated by the statistical representation used in OASES. Long-range transmissions around 900 Hz are found to be more sensitive to the sea-ice roughness than the elastic parameters.

The Arctic Ocean is ringed by shallow areas along the continental shelves which are home to shore-fast ice throughout much of the winter. These areas are being used for both traditional activities such as fishing and hunting, and increasingly, at least in some locations, resource extraction and related exploration. Environmental monitoring under shore-fast ice in multi-use regions will increasingly be accomplished using autonomous underwater vehicles and non-cabled sensors, both of which will require means for acoustic data telemetry. In this paper the results of an experiment done to test the data rate, range and reliability of acoustic communications in 5-10 m deep water under shore-fast ice near Prudhoe Bay, Alaska are described. The experiment was carried out in late winter of 2014 using Micro-Modem communications equipment operating at 10 kHz at ranges of 1.5 to 5.6 km and data rates (burst) from 80 to 5000 bps. The achievable data rates were dependent on range and water depth, plus the presence of ice keels that blocked the acoustic path in some areas. The feasibility of long baseline (LBL) navigation at these ranges is also confirmed by the results.

Acoustic transmissions from shallow sources in the Arctic Ocean can propagate several hundred kilometers due to the presence of an Arctic acoustic duct. Receptions of these long-range transmissions are complex patterns of arrivals which are strongly dependent upon upper ocean sound-speed structure. In addition to measuring sound-speed parameters, gliders equipped with acoustic recorders can measure these arrivals and can complement moored receptions, providing data at many ranges with respect to the moored sources. There is a higher degree of uncertainty in glider position compared with moored receivers, but localization can be improved in post-processing with enhanced acoustic predictability. Two acoustic Seagliders were deployed for a short pilot study in late Summer 2016 in the vicinity of an array of acoustic tomography sources with frequencies on the order of 250 Hz in the Arctic Ocean in anticipation of a longer deployment in Summer 2017. Source receptions recorded on the gliders are compared with acoustic predictions based on sound-speed profiles generated from environmental data collected on the gliders themselves to begin to understand the predictability of transmissions received on gliders in an Arctic environment. Acoustic predictions are analyzed for receptions both within the acoustic duct and at depths exceeding the acoustic duct.

Some forecasts show surface pH in the Arctic dropping from 8.1 to 7.6 over the next 100 years. This substantial decrease may cause changes in acoustic transmission at frequencies where the pH-dependent borate absorption plays a role, below about 5 kHz. In many cases, upward refraction of sound in the near-isothermal Arctic waters causes ice or surface scattering effects to dominate transmission. However, recent observations in the Canada Basin show that 700 and 900-Hz sound can be fully ducted beneath the Pacific Summer Water, with no ice interaction, and be detectable over distances of a few hundred kilometers. In this situation, the received signal level is controlled largely by cylindrical spreading and absorption. Here, for a wide band of frequencies, the effects of probable pH reductions and reduced absorption are investigated using a few models of pH depth profiles. There is potential for increased signal levels of 5 dB or more for 200-km propagation if the duct waters have significantly reduced pH.