Nested Autonomy for Distributed Ocean Sensing 20.6 Application Examples 467
Part B | 20.6
paradigm enables fully autonomous adaptation of the
mobile network nodes to the environmental and tactical
picture, collaborative target event tracking by multiple
platforms, and safe and efficient operation in uncharted
environments without the need for re-programming.
Once deployed, the entire network is operated using
only the DCCL messages for communication between
nodes and human operators for changing mission objectives and platform states. The Goby-DCCL is interfaced
to the MOOS-IvP platform autonomy by the process
pAcommsHandler, as shown in Fig. 20.6, providing the
following capabilities and properties to the autonomy
system:
Highly portable with most of the software being
hardware independent, with generic message handling all the way down to the physical modem
driver.
The DCCL encoding/decoding provides highly efficient data compression through a user-defined message composition with arbitrary value intervals and
resolution.
Dynamic queuing allows for high-priority messages
to move to the head of the queue, with the priority of
less time critical low-priority messages such as status reports increasing with time. This ensures that
the message queue not be saturated by high-priority,
short-time validity messages such as track reports.
User defined time division multiple access (TDMA)
communication scheduling, either using a centralized polling scheme, a fixed slotted scheme, or
a dynamic self-discovering slotted scheme.
20.5 On-Board, Real-Time Signal Processing
A key to the autonomous, adaptive sampling of chemical, biological, physical, or acoustic fields in the ocean
is an efficient on-board implementation of a data analysis package that allows for real-time feedback to
the platform control, allowing for the sensor-adaptive
autonomous control of the platforms. MOOS-IvP provides a very effective infrastructure for achieving this
due to its modular structure and well-defined communication infrastructure. Thus, as illustrated in Fig. 20.6,
separating the signal processing chain into a sequence
of MOOS processes will allow each step in the processing to not only take advantage of the navigation
information available in the MOOSDB, but also have
the possibility of providing feedback to the Helm for
optimizing the processing performance. This real-time,
closed-loop feedback is the key enabler of environmentally adaptive sampling by making processed event
data immediately available to the autonomous control.
Also, it allows the data processing to take advantage
of information arriving from other network nodes and
published in the MOOSDB, enabling more effective
collaborative sampling when the communication environment allows.
20.6 Application Examples
In a collaborative effort between the NATO Undersea
Research Centre (NURC), MIT, Woods Hole Oceanographic Institution (WHOI), and the Naval Undersea
Warfare Center (NUWC), a series of experiments were
carried out between 2008 and 2010, with the primary
objectives being the demonstration of the performance
of a network of underwater vehicles as receiver platforms for multistatic active sonar tracking, and their
communication and control networking. In these experiments, up to seven AUVs were deployed, six of which
were equipped with towed hydrophone arrays and operating in a common underwater acoustic communication
network [20.2]. These experiments provided a unique
opportunity for testing and demonstrating the performance of the nested autonomy control paradigm under
realistic underwater communication constraints. The
experiments were designed, in addition to the multistatic acoustic data collection, to allow the demonstration of fully autonomous oceanographic mapping
and adaptive autonomous behaviors for optimal acoustic sensing and communication. They also provided
a comprehensive testbed for the MOOS-IvP platform
autonomy in general.
20.6.1 Unified Command,
Communication,
and Control Infrastructure
The GLINT’08, ’09, and ’10 experiments were carried out as part of a joint research project (JRP) on
undersea sensing network technology (NURC project
4G4), involving NURC, MIT, WHOI, NUWC, and
several Italian organizations. The experiments had several scientific objectives, relating both to the sensing
Part B | 20.6
paradigm enables fully autonomous adaptation of the
mobile network nodes to the environmental and tactical
picture, collaborative target event tracking by multiple
platforms, and safe and efficient operation in uncharted
environments without the need for re-programming.
Once deployed, the entire network is operated using
only the DCCL messages for communication between
nodes and human operators for changing mission objectives and platform states. The Goby-DCCL is interfaced
to the MOOS-IvP platform autonomy by the process
pAcommsHandler, as shown in Fig. 20.6, providing the
following capabilities and properties to the autonomy
system:
Highly portable with most of the software being
hardware independent, with generic message handling all the way down to the physical modem
driver.
The DCCL encoding/decoding provides highly efficient data compression through a user-defined message composition with arbitrary value intervals and
resolution.
Dynamic queuing allows for high-priority messages
to move to the head of the queue, with the priority of
less time critical low-priority messages such as status reports increasing with time. This ensures that
the message queue not be saturated by high-priority,
short-time validity messages such as track reports.
User defined time division multiple access (TDMA)
communication scheduling, either using a centralized polling scheme, a fixed slotted scheme, or
a dynamic self-discovering slotted scheme.
20.5 On-Board, Real-Time Signal Processing
A key to the autonomous, adaptive sampling of chemical, biological, physical, or acoustic fields in the ocean
is an efficient on-board implementation of a data analysis package that allows for real-time feedback to
the platform control, allowing for the sensor-adaptive
autonomous control of the platforms. MOOS-IvP provides a very effective infrastructure for achieving this
due to its modular structure and well-defined communication infrastructure. Thus, as illustrated in Fig. 20.6,
separating the signal processing chain into a sequence
of MOOS processes will allow each step in the processing to not only take advantage of the navigation
information available in the MOOSDB, but also have
the possibility of providing feedback to the Helm for
optimizing the processing performance. This real-time,
closed-loop feedback is the key enabler of environmentally adaptive sampling by making processed event
data immediately available to the autonomous control.
Also, it allows the data processing to take advantage
of information arriving from other network nodes and
published in the MOOSDB, enabling more effective
collaborative sampling when the communication environment allows.
20.6 Application Examples
In a collaborative effort between the NATO Undersea
Research Centre (NURC), MIT, Woods Hole Oceanographic Institution (WHOI), and the Naval Undersea
Warfare Center (NUWC), a series of experiments were
carried out between 2008 and 2010, with the primary
objectives being the demonstration of the performance
of a network of underwater vehicles as receiver platforms for multistatic active sonar tracking, and their
communication and control networking. In these experiments, up to seven AUVs were deployed, six of which
were equipped with towed hydrophone arrays and operating in a common underwater acoustic communication
network [20.2]. These experiments provided a unique
opportunity for testing and demonstrating the performance of the nested autonomy control paradigm under
realistic underwater communication constraints. The
experiments were designed, in addition to the multistatic acoustic data collection, to allow the demonstration of fully autonomous oceanographic mapping
and adaptive autonomous behaviors for optimal acoustic sensing and communication. They also provided
a comprehensive testbed for the MOOS-IvP platform
autonomy in general.
20.6.1 Unified Command,
Communication,
and Control Infrastructure
The GLINT’08, ’09, and ’10 experiments were carried out as part of a joint research project (JRP) on
undersea sensing network technology (NURC project
4G4), involving NURC, MIT, WHOI, NUWC, and
several Italian organizations. The experiments had several scientific objectives, relating both to the sensing