Objective: The primary research objective is to develop a
Semi-Autonomous Underwater Vehicle for Intervention Missions (SAUVIM).
Unlike the fly-by autonomous underwater vehicles (AUV), SAUVIM will have
a manipulator work package. It will require an advanced control system and
a precise sensory system to maintain high accuracy in station keeping and
navigation.
Background: Most intervention missions, including underwater
plug/unplug, construction & repair, cable streaming, mine hunting, and
munitions retrieval- require physical contact with the surroundings in
the unstructured, underwater environment. Such operations always increase
the level of risk and present more difficult engineering problems than
fly-by and non-contact type operations. For these intervention operations,
the vehicle requires a dexterous robotic manipulator; thus the overall system
becomes a high degree-of-freedom (dof), multi-bodied system from the coupling
effects of the high degree of accuracy even in the presence of unknown,
external disturbances, i.e. undersea currents. All these issues present
very complex engineering problems that have hindered the development of
AUVs for intervention missions. Currently, the state-of-the-art in machine
intelligence is insufficient to create a vehicle of full autonomy and
reliability, especially for intervention missions.
Development: Five major components: Adaptive, Intelligent Motion
Planning; Automatic Object Ranging and Dimensioning; Intelligent Coordinated
Motion/Force Control; Predictive Virtual Environment; and SAUVIM Design.
SAUVIM Design (SD) This task is the main objective of the SAUVIM project
for Phase 1. It is an effort to design and develop efficient, reliable
hardware/software architectures of SAUVIM. Due to the immense demand of
this task, it is divided into five sub-tasks, which are Reliable,
Distributed Control (RDC), Mission Sensor Package (MSP), Hydrodynamic
Drag Coefficient Analysis (HDCA), Mechanical Analysis and Fabrication (MAF),
and Mechanical-Electrical Design (MED).
The goal of RDC is to develop a reliable and efficient computing
architecture for signal and algorithmic processing of the entire SAUVIM
system. The proposed system is a multi-processor system based on a 6U VMEbus
and the VxWorks real-time operating system. This system is capable of high
processing throughput and fault tolerance. Currently the system consists
of two VMEbuses, which are the navigation control system and the manipulator
control system. The main VMEbus, or the navigation control system, has two
Motorola M68060 CPU boards, a multi-port RS232 interface board, and an I/O
board with a Pentium MMX processor based PC104+ board, which is connected
via a RS232 port. The navigation control system handles the communication,
supervision, planning, low-level control, self-diagnostics, video imaging,
etc. The data exchange between the two CPUs is conducted via shared memory.
The second VMEbus, or the manipulator control system, has one Motorola M68040
CPU and an I/O board. Two PC104 boards are connected serially to this CPU.
The manipulator control system is independent and dedicated to the
manipulator control. Many of the hardware components have been tested and
are being interface with its respective software systems. Various
optimization changes have been implemented to minimize communication
and computation. This development will continue throughout the vehicle's
development process.
The objective of the MSP is to provide semi-continuous records of
SAUVIM water depth, temperature, conductivity, computed salinity,
dissolved oxygen, magnetic signature of the seafloor, pH and turbidity
during the survey mode. In the intervention mode, the MSP also provides
compositional parameters at a selected seafloor target, including pumped
samples from submarine seeps or vents. The MSP is an independent system
with its own PC 104 CPU and its own power supply residing in a separate
pressure vessel. All of the sensors have been purchased, and an initial
field test at the Loihi Seamount has been conducted. Continual tests are
being conducted to optimize the scientific sensor data-gathering
capabilities.
The HDCA is used to determine the hydrodynamic coefficients via a
numerical solution of full Navier-Stokes equations using PHOENICS, a
commercial computational fluid dynamics (CFD) code. Initial results
from the PHOENICS software have produced mixed results. The current
vehicle fairing has produced a drag coefficient of 0.40; however,
it has not yet been verified. Other CFD software and model testing is
being conducted to verify the drag coefficient results before the
implementation of the vehicle fairing on SAUVIM.
The MAF has three objectives. Its primary goal is to design and f
abricate composite pressure vessels with end caps and connector openings
for full ocean depths taking stress, buckling, Hygrothermal effects, and
fatigue analysis into account; and its two secondary goals are to design
and fabricate the SAUVIM fairing and to analyze the SAUVIM frame. A thorough
analysis and comparison of the Ti-6Al4V, AS4/Epoxy, and AS4/PEEK pressure
vessels manifest the advantage of composite materials in reduction of weight,
size and strength. Using these results, a scaled model prototype using
AS4/PEEK has been fabricated and tested. A l/3 sized prototype is being
fabricated and will also be tested. For the shallow water vehicle test,
a full-sized, fiberglass pressure vessel with aluminum end caps have been
manufactured and tested. These vessels are being used to determine the final
hardware layout. The aluminum frame has been designed and fabricated. A
full-ocean depth pressure vessel of AS4/PEEK has been developed and is in
its testing phase. The initial fairing analysis has been developed and
expanded. Fairing optimizations are being considered.
The MED is the integration of the mechanical and electrical components
for SAUVIM. First, the design specifications were established for the
fairing, frame, instrument pressure vessels, buoyancy systems, mission
sensor, passive arm and robotic manipulator tasks. Second, after scrutinizing
review of SAUVlM's major components - i.e. sensors, actuators and
infrastructure -in terms of power consumption, compatibility, weight
distribution, buoyancy distribution, hydrodynamic effects and task
effectiveness, all major components have been purchased. Technical drawings
of the vehicle frame, fairing, and related sub-structures have been
completed. Many of the mechanical and electrical components have been
fabricated and are being integrated with the overall electrical layouts.
|