autonomous ocean-sampling networks - IEEE Xplore

3 downloads 88294 Views 32KB Size Report
Compact and affordable, autonomous underwater vehicles. (AUVs) supported by ... chitectures involve a cluster of platforms linked to a server or gateway node .... recently resurrected Iridium system and low-cost, dedicated, store-and-forward ...
IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. 26, NO. 4, OCTOBER 2001

421

Guest Editorial Autonomous Ocean-Sampling Networks

T

HE autonomous ocean-sampling network (AOSN) paradigm provides a nested ocean observation capability through the coordinated control of many, mobile, networked, sensor platforms (Curtin et al., 1993). In its most comprehensive form, the concept closely couples observational tools with modeling capabilities to reduce error in ocean state estimation. Concerted efforts over the last half-decade have developed and demonstrated key elements of the AOSN. This special issue provides a representative status report of progress toward the AOSN goal. All references below are to papers in this issue. Compact and affordable, autonomous underwater vehicles (AUVs) supported by a communication and power infrastructure, comprise the basic observation system. To span the dynamic range for observing the spatial and temporal scales of ocean processes, a variety of vehicle capabilities have been developed. Slow-moving, buoyancy-driven vehicles are employed for monitoring large and meso-scale processes. These vehicles, also called gliders, were largely conceptual when the AOSN initiative was first formulated, but have now been fielded by a number of developers. Eriksen et al. in an initial report on Seaglider, a vehicle with low–drag hull form and glide slopes ranging from 0.2 to 3 designed for missions of several thousand kilometers and many months, describe remote command and near real time data telemetry during Puget Sound field trials. Sherman et al. in an initial report on Spray, a glider with operating speeds of 20–30 cm/s and ranges up to 6000 km, describe observations of internal waves and tides in Monterey Bay canyon. Webb et al. in a report on Slocum, a glider that extends the operational range to 40000 km by harvesting its propulsive energy from the vertical thermal gradient in the temperate and tropical ocean, describe performance during initial field trials. All these gliders can be launched manually by two people from a small boat, and can be programmed or commanded to return to a desired location for retrieval at the completion of a mission. Faster, more maneuverable, propeller-driven vehicles enable observation of smaller scale, rapidly evolving processes. Smith et al. report on Morpheus design and results from mine counter measure missions. Morpheus extends modularity to a new level in small (two-person portable), propeller-driven platforms with mechanical and electrical assembly of interoperable, molded plastic hull modules, enabling rapid reconfiguration from 1 to 3 m in length depending on the mission. Typical network architectures involve a cluster of platforms linked to a server or gateway node vehicle. Marco and Healey report on architecture and navigation accuracy for ARIES, an early prototype of a mobile, gateway network node. The practical implementation of an AOSN depends on the availability of network-class vehicles:

Publisher Item Identifier S 0364-9059(01)10903-9.

platforms that are affordable both in initial unit cost and in logistical support for deployment. Network-class propeller-driven and buoyancy-driven vehicles, such as those described in this issue, have been proven in numerous field experiments reported here and elsewhere. Some have matured to the point of commercial availability, although the economies of scale have yet to be realized. To enable extended deployment for propeller-driven vehicles, a variety of experimental docking systems have been fielded. Docking stations for re-fueling can be fixed or mobile. Fixed stations can be on the bottom or located in the water column on moorings. Two general purpose mooring designs, deep and shallow, have been developed and tested to support AUV operations in AOSNs (Frye et al.). A fixed bottom station has been deployed at LEO-15 off New Jersey, where a cable provides shore connectivity (Stokey et al.). Future mobile docking stations, essentially tanker AUVs, will capitalize on fixed docking system research. Singh et al. describe an initial omni-directional docking method with acoustic terminal homing and novel mechanical and electrical connection mechanisms. This method continues to be simplified and refined. Feezor et al.provides details on an inductive coupler for noninvasive power and data transfer, as well as on electromagnetic homing, which provides an alternative to acoustics with advantages in some applications. Terminal homing with optical systems has also proven successful. Unidirectional docking methods, like runways in aviation, trade the advantages of simplified vehicle manipulation once docked for a more constrained approach path. Stokey et al.report on oriented docking using REMUS. Localization and mating are only part of a complete docking cycle, which includes efficient power management and adequate energy supply. Bradley et al. investigate the issues involved in battery chemistry, pack management and in-situ charging for long term deployments at remote sites, often with low ambient temperatures. The fuel at docking stations can be a pre-positioned cache of batteries, fuel cells or thermal sources and/or it can be harvested locally from the kinetic, thermal or chemical energy in the environment. For long term deployments with modest duty cycles, harvesting is attractive. Taylor et al. describe “eel” generators that use wake vortices in the flow past a bluff body to strain piezoelectric elements, converting mechanical energy to electrical power. Precise undersea navigation with minimal use of external aides (GPS, acoustic transponders or electromagnetic sources) is critical for long transits and accurate estimation of spatial gradients. The heart of such navigation is an inertial system such as that described by Grenon et al., which uses an extended Kalman filter to fuse ring-laser gyros, accelerometers, Doppler velocity log and occasional GPS fixes. Concurrent mapping and localization provides a complementary approach to inertial guidance, enabling an AUV to build a feature-based map of

0364–9059/01$10.00 © 2001 IEEE

422

an unknown environment while using that map to navigate. Leonard and Feder report on a new computationally efficient method for subdividing the domain to address the practical scaling issues limiting the use of this approach. A combination of acoustic, line-of-sight, and over-the horizon direct radio, and satellite communications will link observation to modeling systems and enable coordinated control of an AOSN. Command of and data retrieval from AUVs across transcontinental distances has been demonstrated using a combination of internet and nearshore radio modem connections. A number of undersea acoustic modems are now available for point-to-point communication. Both multiple access interference and inter-symbol interference due to multipath propagation ultimately determine acoustic communication throughput between any two platforms in an AOSN. Reliable signal detection in noisy environments without excessive false alarms is a prerequisite to further signal processing. Preisig and Johnson report on high performance detection algorithms that exploit the structure of frequency hopped signals interleaved with quiescent bands. Freitag et al.analyze two code-division spread-spectrum signaling methods for multiple-access communication: phase-modulated, direct sequence spread spectrum, and noncoherent, frequency-hopped spread spectrum. They examine performance trade offs in shallow water where the limitations in the temporal coherence of the channel must be balanced against the ability to resolve multipath. Tsimenidis et al.evaluate an adaptive-array receiver structure combining direct-sequence code-division multiple access and spatial diversity to achieve reliable, low data rate, multiuser communication in an asynchronous shallow water network. In an analysis of a number of multiuser, detection receiver structures employing adaptive decision feedback equalization and spatial diversity, Yeo et al.report best performance in the North Sea using recursive successive interference cancellation. For a Doppler spread channel with a given number of degrees of freedom, Eggen et al. derive the equalization filter length limits for the stability of a coherent receiver. The above-sea radio link, while potentially of much greater bandwidth than the undersea acoustic channel, typically relies on costly infrastructure (e.g., satellite systems) not targeted for autonomous, low power marine users. Opportunities for satellite communication links beyond ARGOS are continually evolving. Low earth orbit options now include the recently resurrected Iridium system and low-cost, dedicated, store-and-forward micro-satellites. These options are currently being pursued. Here, Gamache and Wetherby report on the design and performance of a prototype system for remote, low power applications using existing, C-band, geosynchronous satellites. Adaptive control of an AOSN is inherently hierarchical. Building on individual vehicle controllers, groups of vehicles will be managed by a “server” vehicle providing a more global view of the state of the environment and mission constraints. The number of groups and membership in each group are dynamic in response to sensor observations, fault detection, and mission planning. First generation, adaptive networks of gliders for mesoscale ocean sampling are ready to be deployed today. With this objective in mind, Leonard and Graver derive

IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. 26, NO. 4, OCTOBER 2001

feedback control laws to make glider motion robust to disturbances and uncertainty. Much can be learned about intelligent control strategies through realistic simulation. Phoha et al. report on a simulation experiment that demonstrates multistage inferencing and decision/control strategies for the evolution of group behavior in response to evolving ocean and operational dynamics. AOSN assets can be viewed as autonomous agents that follow protocols to create an effective organization to execute a mission. Turner and Turner describe a two-level, self-organizing, control architecture that allows an AOSN to be efficient and flexible in response to change, and they demonstrate its behavior when agents fail. The platforms in an AOSN are only as useful as the sensors they carry. Many compact, low power, high precision sensors are now being developed for AUVs. Four are reported here. A major advantage of an AOSN is that a diversity of sensors can be distributed simultaneously in space and their data fused to provide both synoptic spatial gradients and multivariate vectors for process and target signature identification. Kaltenbacher et al.describe a high sensitivity, spectrophotometric chemical sensor, the spectral elemental analysis system (SEAS), that can be tuned to sense various elements. SEAS great sensitivity derives from an optical absorption cell that is a flexible tube several meters long acting as a liquid core waveguide. Samson et al. report on the shadowed image-particle profiling and evaluation recorder (SIPPER), an instrument capable of imaging micron to centimeter scale particles insitu with digital output compatible with automated counting, sizing and identification. Utilizing combinations of new chemical and particle sensors, AOSN missions will include Lagrangian chemical plume and plankton patch tracking and evolution. Schock et al.reports on the performance of a new sonar with a line array of sources and a planar receive array designed to scan for objects buried in the seaflooor. Optimal for near-bottom-following AUV operation, the receive array utilizes nearfield focusing which allows the sonar to operate near the seabed where target images have the highest signal-to-noise and resolution. Another approach to buried object detection lies in expanding the aperture with bior multi-static configurations using a number of vehicles. Edwards et al. examine the feasibility of combining seabed scattering data from consecutive pings of a fixed parametric source to form a bistatic synthetic aperture for target localization and imaging with an AUV-based receiver, and assess the performance tradeoffs associated with different levels of processing. An intentionally grounded AUV provides an ideal platform for a long term, extended aperture, passive sonar with the flexibility of re-positioning during a mission to further investigate sources of interest. Glegg et al. report on the successful development of an umbrella array that operates in this mode. The AOSN was conceived as a tool for mapping ocean properties and providing targeted observations to validate and improve the forecasting skill of models. In four dimensional state estimation, both sensor inaccuracy and spatial-temporal aliasing contribute to errors. Willcox et al. describe a graphical approach to designing an AUV survey that meets prescribed error tolerances for specific ocean processes. In nature, a spectrum of processes occurs simultaneously. Zhang et al.report on a parametric tool for designing an AUV process classifier based on the

IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. 26, NO. 4, OCTOBER 2001

relationship between observations from a moving platform and the frequency-wavenumber spectrum of the surveyed process. A case study distinguishes internal waves from deep convection in the Labrador Sea. To date, a number of experiments (e.g., Haro Strait, Massachusetts Bay, LEO-15 off New Jersey) have demonstrated adaptive AUV sampling guided by prognostic ocean models. Carder et al. describe a similar effort on the West Florida Shelf. The cueing of AUV-based subsurface sampling by surface remote sensing can effect efficient gradient tracking and full volumetric assessment of surface features. If signatures are unique, remotely sensed fields can be projected using such data. Here, Dhanak et al. examine sub surface current radar measurements. AOSN is a pioneering effort in concept, components and system integration. Since outlining this paradigm in 1993, we are inspired and gratified by the progress that this collection of papers represents. AOSN has had an impact beyond the original scope of oceanographic sampling. The component capabilities

423

developed have been adapted to applications ranging from mine countermeasures to oil industry surveys, and are changing how we observe and predict the ocean. Thirty of the papers contained in this issue report on research funded by the Office of Naval Research (ONR). The “concerted effort” has been a productive partnership between the research community and colleagues at ONR who continue to contribute in many tangible and intangible ways. THOMAS B. CURTIN, Guest Editor Office of Naval Research Code 322OM Arlington, VA 22217 JAMES G. BELLINGHAM, Guest Editor Monterey Bay Aquarium Research Institute Department of Engineering Moss Landing, CA95039

Thomas B. Curtin received the B. S. degree from Boston College, Boston, MA, the M. S. degree from Oregon State University, Corvallis, and the Ph.D degree from the University of Miami, Miami, FL , in 1963, 1969, and 1979, respectively. Between 1970 and 1973, he was with the Peace Corps, Malaysia. Betweem 1980 and 1984, he was an Assistant Professor at the North Carolina State University, Raleigh. Since 1984, he has managed programs in physical oceanography, Arctic sciences, and ocean modeling and prediction, at the Office of Naval Research. He has developed major initiatives in surface gravity waves, Arctic leads, sea-ice mechanics, Arctic and marine mammal acoustics, deep ocean convection and autonomous sampling networks. Dr. Curtin has led over 25 oceanographic cruises in mid-latitude and equatorial Atlantic and Pacific Oceans, the Ross Sea, and the South China Sea. He has been awarded the U. S. Navy Meritorious and Superior Civilian Service Medals.

James G. Bellingham received the a M.S., and Ph.D. degrees in physics, from the Massachusetts Institute of Technology (MIT), Cambridge, MA, in 1984, and 1988, respectively. He spent the last thirteen years developing Autonomous Underwater Vehicles (AUV), first as Manager of the MIT Sea Grant College Program AUV Laboratory and, more recently, as the Director of Engineering at the Monterey Bay Aquarium Research Institute (MBARI). He has worked in areas ranging from vehicle design, to high-level control, to field operations. His primary contribution is the creation of the Odyssey class of AUVs with which he has led operations in areas as remote as the Arctic and Antarctic. In the area of high-level control, he developed state-configured layered control, a new methodology for handling mission-level control of autonomous vehicles. Dr. Bellingham is a founder and member of the Board of Directors of Bluefin Robotics Corporation, a leading manufacturer of AUVs for the military, commercial, and scientific markets.