Preliminary experiments in model-based thruster control forunderwater vehicle positioning

This paper reports comparative experiments with two novel and one conventional thrust control algorithms for the unsteady (transient) control of thrust generated by conventional bladed-propeller marine thrusters. First, comparative experiments with three different thrust control algorithms over a wide range of unsteady operating conditions suggest that model-based control algorithms offer transient thrust-control performance superior to that of their nonmodel-based counterpart. Second, hybrid simulations combining actual real-time experimental thruster responses with simulated one-dimensional real-time vehicle dynamics suggest that model-based thrust control algorithms offer vehicle position control superior to that of its nonmodel-based counterpart     

Hydrothermal activity on the southern Mid-Atlantic Ridge: Tectonically- and volcanically-controlled venting at 4–5°S

We report results from an investigation of the geologic processes controlling hydrothermal activity along the previously-unstudied southern Mid-Atlantic Ridge (3–7°S). Our study employed the NOC (UK) deep-tow sidescan sonar instrument, TOBI, in concert with the WHOI (USA) autonomous underwater vehicle, ABE, to collect information concerning hydrothermal plume distributions in the water column co-registered with geologic investigations of the underlying seafloor. Two areas of high-temperature hydrothermal venting were identified. The first was situated in a non-transform discontinuity (NTD) between two adjacent second-order ridge-segments near 4°02′S, distant from any neovolcanic activity. This geologic setting is very similar to that of the ultramafic-hosted and tectonically-controlled Rainbow vent-site on the northern Mid-Atlantic Ridge. The second site was located at 4°48′S at the axial-summit centre of a second-order ridge-segment. There, high-temperature venting is hosted in an  18 km² area of young lava flows which in some cases are observed to have flowed over and engulfed pre-existing chemosynthetic vent-fauna. In both appearance and extent, these lava flows are directly reminiscent of those emplaced in Winter 2005−06 at the East Pacific Rise, 9°50′N and reference to global seismic catalogues reveals that a swarm of large (M 4.6−5.6) seismic events was centred on the 5°S segment over a  24 h period in late June 2002, perhaps indicating the precise timing of this volcanic eruptive episode. Temperature measurements at one of the vents found directly adjacent to the fresh lava flows at 5°S MAR (Turtle Pits) have subsequently revealed vent-fluids that are actively phase separating under conditions very close to the Critical Point for seawater, at  3000 m depth and 407 °C: the hottest vent-fluids yet reported from anywhere along the global ridge crest.     

The Hawaii-2 Observatory

A permanent deep ocean scientific research facility-the Hawaii-2 Observatory (H2O)-was installed on the retired HAW-2 commercial submarine telephone cable in September 1998. H2O consists of a seafloor submarine cable termination and junction box in 5000 m of water located halfway between Hawaii and California. The H2O infrastructure was installed from a large research vessel using the Jason ROV and standard over-the-side gear. The junction box provides two-way digital communication at variable data rates of up to 80 kbit/s using the RS-422 protocol and a total of 400 W of power for both junction box systems and user equipment. Instruments may be connected by an ROV to the junction box at 8 wet-mateable connectors. The H2O junction box is a "smart" design, which incorporates redundancy to protect against failure and allows full control of instrument functionality from shore. Initial instrumentation at the H2O site includes broad-band seismometer and hydrophone packages     

Docking for an autonomous ocean sampling network

In this paper, we examine the issues associated with docking autonomous underwater vehicles (AUVs) operating within an Autonomous Ocean Sampling Network (AOSN). We present a system based upon an acoustic ultrashort baseline system that allows the AUV to approach the dock from any direction. A passive latch on the AUV and a pole on the dock accomplish the task of mechanically docking the vehicle. We show that our technique for homing is extremely robust in the face of the two dominant sources of error-namely the presence of currents and the presence of magnetic anomalies. Our strategy for homing is independent of the initial bearing of the dock to the AUV, includes a method for detecting when the vehicle has missed the dock, and automatically ensures that the AUV is in a position to retry homing with a greater chance of success. Our approach is seen to be extremely successful in homing the vehicle to the dock, mechanically attaching itself to the dock, aligning inductive cores for data and power transfer, and undocking at the start of a fresh mission. Once the AUV is on the dock, we present a methodology that allows us to achieve the complex tasks with ensuring that the AUV is securely docked, periodically checking vehicle status, reacting to a vehicle that requires charging, tracking it when it is out on a mission, archiving and transmitting via satellite the data that the AUV collects during its missions, as well as providing a mechanism for researchers removed from the site to learn about vehicle status and command high-level missions. The dock is capable of long-term deployments at a remote site while respecting the constraints - low power, small size, low computational energy, low bandwidth, and little or no user input - imposed by the amalgamation of acoustic, electronic and mechanical components that comprise the entire system     

The influence of thruster dynamics on underwater vehicle behaviorand their incorporation into control system design

A nonlinear parametric model of a torque-controlled thruster is developed and experimentally confirmed. The model shows that the thruster behaves like a sluggish nonlinear filter, where the speed of response depends on the commanded thrust level. A quasi-linear analysis which utilizes describing functions shows that the dynamics of the thruster produce a strong bandwidth constraint and a limit cycle, which are both commonly seen in practice. Three forms of compensation are tested, utilizing a hybrid simulation combining an instrumented thruster with a real-time mathematical vehicle model. The first compensator, a linear lead network, is easy to implement and greatly improves performance over the uncompensated system, but does not perform uniformly over the entire operating range. The second compensator, which attempts to cancel the nonlinear effect of the thruster, is effective over the entire operating range but depends on an accurate thruster model. The final compensator, an adaptive sliding controller, is effective over the entire operating range and can compensate for uncertainties or the degradation of the thruster     

Teleprogramming for subsea teleoperation using acousticcommunication

This paper considers the performance of subsea intervention tasks from an unmanned untethered submersible while using acoustic communications. It is argued that the low bandwidth and high delay imposed by acoustic modems makes it unwise to adopt conventional teleoperation techniques and a system is presented which permits subsea teleoperative tasks to be carried out using such limited communication resources. The described implementation employs active techniques to assist the operator both in performing actions and in recovering from those problems which will inevitably occur during real-world interaction. It provides the operator with both simulated and real visual imagery and is designed to adapt dynamically to changing bandwidth and computational resources. Experiments are described in which an operator in Philadelphia, PA, controlled a robot manipulator mounted on the JASON underwater vehicle submerged off the Massachusetts coast. All communication over this 500-km distance was via a combination of Internet and a simulated acoustic link. Analysis of the bandwidth requirements showed them to be consistent with those from acoustic subsea networks     

Supervisory control system for the JASON ROV

A prototype supervisory control system for a remotely operated vehicle (ROV) is described and several key elements demonstrated in simulation and in-water tests. This system is specifically designed to fill the needs of JASON, a new ROV under development that will perform scientific tasks on the seafloor to depths of 6000 m. JASON will operate from the ARGO towed imaging platform, which is currently operational. Supervisory control is a paradigm for combined human and computer control. Several key elements of the supervisory control system are presented. These include the closed-loop positioning system based on a high-resolution acoustic navigation system, a monitoring capability for assessing performance and detecting undesirable changes, and an interface that allows the human operator and the computer system to specify the desired vehicle trajectory jointly.     

Robust trajectory control of underwater vehicles

underwater vehicles present difficult control-system design problems due to their nonlinear dynamics, uncertain models, and the presence of disturbances that are difficult to measure or estimate. In this paper, a recent extension of sliding mode control is shown to handle these problems effectively. The method deals directly with nonlinearities, is highly robust to imprecise models, explicitly accounts for the presence of high-frequency unmodeled dynamics, and produces designs that are easy to understand. Using a nonlinear vehicle simulation, the relationship between model uncertainty and performance is examined. The results show that adequate controllers can be designed using simple nonlinear models, but that performance improves as model uncertainty is decreased and the improvements can be predicted quantitatively.     

Development, comparison, and preliminary experimental validation ofnonlinear dynamic thruster models

This paper examines dynamical models for bladed-propeller-type marine thrusters. Previously reported thruster dynamics models are reviewed, and a simplified model is proposed. Experimental testing of both the transient and steady-state performance of a marine thruster corroborates previously reported data, validates the simplified thruster model, and raises new questions. A companion paper reports preliminary experiments in the design and testing of thrust controllers which incorporate the dynamical thruster models described in the present paper