A Networked Robotic System and its Use in an Oil Spill Monitoring Exercise
Eloi Pereira, Pedro Marques da Silva, Clemens Krainer, Christoph M. Meyer, Jose Morgado, Raja Sengupta

Citation
Eloi Pereira, Pedro Marques da Silva, Clemens Krainer, Christoph M. Meyer, Jose Morgado, Raja Sengupta. "A Networked Robotic System and its Use in an Oil Spill Monitoring Exercise". Talk or presentation, 29, September, 2013; Presented at the First International Workshop on the Swarm at the Edge of the Cloud (SEC'13 @ ESWeek), Montreal.

Abstract
We have contributed with a bridging model, akin to the von Neumann model, to the cyber- physical systems literature. This past contribution, named the BigActor model, is only a mathematical model. Here we describe a software system built to explore the value of the mathematics when integrating and controlling a network of robots sensing an environment. Figure 1(b) describes the “physical” components of the system as well as their relative location and connectivity. This particular robot network executes an oil spill monitoring exercise. There is one UAV, 3 ground control stations, 4 drifters that broadcast their position using Automatic Identification System (AIS), and one ship in the system. These are denoted as labeled boxes in Figure 1(b). Figure 1(a) shows the same information but in the physical space, i.e. the GPS coordinates of the vehicles. There are three communication channels. The AIS channel is used by the drifters, the piccolo channels by the UAV and ground stations. The other boxes in the picture are static locations representing areas like the take-off airfield, the oil-spill area, etc. The picture is formally a Bigraph as defined by Milner in [2]. We use it exactly as Milner intended to represent the ubiquitous computing machine, analogous to the von Neumann machine. We control the system by commanding an entity such as the UAV to go from the airfield to the oil spill, or to disconnect from one ground station to another ground station, to represent logically the transfer of control of the asset from one jurisdiction to another (in one of our exercises the the ground station was onboard of a Navy vessel). These controls are formally Bigraph Reaction Rules (BRR). Thus the oil spill monitoring exercise is abstractly a sequence of pictures or Bigraphs, evolving in response to the controls or environmental disturbances like network disconnections due to limited communication range. The “cyber” components controlling the system are BigActors. These are mathematically Actors as defined by Agha and Hewitt bridged to the underlying Bigraph as defined in. BigActors are actors equipped with the ability to observe and control the Bigraph. Figure 2(a) is an image captured from the UAV of the naval vessel and the emulated oil spill. The Portuguese Navy emulated the oil spill by releasing 100 kg of popcorn in the ocean (the yellow patch). The stills have been extracted from the UAV video to verify that the BigActor implementation is correctly able to control the UAV asset through take-off, three ground station handoffs, arrival in the oil spill area, imaging of the spill, connect with the AIS drifters, acquire their data, and deliver it back to base. The creative engineering contributions in the implementation of the BigActor model are an approximation of the BigActor semantics enabling distributed control, control by logical space programming, as opposed to physical space programming, and a process to build networks out of different vehicles built by different teams. Unlike the von Neumann machine, the structure of our ubiquitous computing machine changes as the oil spill exercise executes. These changes of structure are first locally observed by each vehicle in the system. These local observations are then propagated throughout the overall system, flooding the information in a distributed manned. Our BigActor implementation has a distributed Bigraph synthesis protocol and se- mantics to enable coordinated control with decentralized information. The implementation supports control by logical space programming because the distributed Bigraph synthesis pro- tocol provides each BigActor with a Bigraph abstraction of the whole system, enabling the actor to control at the Bigraph level. The Bigraph is a logical, not physical, space abstraction of the real world. Protocols in the implementation handle the binding of logical space to phys- ical space, ensuring Bigraph Reaction Rules actually have a physical effect. This is performed using the Robot Operating System (ROS) as the underlying middleware. We have created a ROS abstraction of each robot in the network called the ros vehicle and defined a set of libraries that make a robot into a ros vehicle. To put a new vehicle into the robot networks one should implement or install these libraries. They make the new vehicle into a ros vehicle, and thereby a node in the network. The network control is at the BigActor level. This is implemented using the Scala Actor Library running over ROS using ROSJava. The overall system’s architecture is presented in Figure 2(b).

Electronic downloads

Citation formats  
  • HTML 
    Eloi Pereira, Pedro Marques da Silva, Clemens Krainer,
Christoph M. Meyer, Jose Morgado, Raja Sengupta. <a
href="http://www.terraswarm.org/pubs/123.html"><i>A
Networked Robotic System and its Use in an Oil Spill
Monitoring Exercise</i></a>, Talk or
presentation,  29, September, 2013; Presented at the <a
href="http://www.terraswarm.org/conferences/13/swarm/index.htm"
>First International Workshop on the Swarm at the Edge of
the Cloud (SEC'13 @ ESWeek)</a>, Montreal.
  • Plain text 
    Eloi Pereira, Pedro Marques da Silva, Clemens Krainer,
Christoph M. Meyer, Jose Morgado, Raja Sengupta. "A
Networked Robotic System and its Use in an Oil Spill
Monitoring Exercise". Talk or presentation,  29,
September, 2013; Presented at the <a
href="http://www.terraswarm.org/conferences/13/swarm/index.htm"
>First International Workshop on the Swarm at the Edge of
the Cloud (SEC'13 @ ESWeek)</a>, Montreal.
  • BibTeX 
    @presentation{PereiraMarquesdaSilvaKrainerMeyerMorgadoSengupta13_NetworkedRoboticSystemItsUseInOilSpillMonitoringExercise,
    author = {Eloi Pereira and Pedro Marques da Silva and
              Clemens Krainer and Christoph M. Meyer and Jose
              Morgado and Raja Sengupta},
    title = {A Networked Robotic System and its Use in an Oil
              Spill Monitoring Exercise},
    day = {29},
    month = {September},
    year = {2013},
    note = {Presented at the <a
              href="http://www.terraswarm.org/conferences/13/swarm/index.htm"
              >First International Workshop on the Swarm at the
              Edge of the Cloud (SEC'13 @ ESWeek)</a>, Montreal.},
    abstract = {We have contributed with a bridging model, akin to
              the von Neumann model, to the cyber- physical
              systems literature. This past contribution, named
              the BigActor model, is only a mathematical model.
              Here we describe a software system built to
              explore the value of the mathematics when
              integrating and controlling a network of robots
              sensing an environment. Figure 1(b) describes the
              âphysicalâ components of the system as well as
              their relative location and connectivity. This
              particular robot network executes an oil spill
              monitoring exercise. There is one UAV, 3 ground
              control stations, 4 drifters that broadcast their
              position using Automatic Identification System
              (AIS), and one ship in the system. These are
              denoted as labeled boxes in Figure 1(b). Figure
              1(a) shows the same information but in the
              physical space, i.e. the GPS coordinates of the
              vehicles. There are three communication channels.
              The AIS channel is used by the drifters, the
              piccolo channels by the UAV and ground stations.
              The other boxes in the picture are static
              locations representing areas like the take-off
              airfield, the oil-spill area, etc. The picture is
              formally a Bigraph as defined by Milner in [2]. We
              use it exactly as Milner intended to represent the
              ubiquitous computing machine, analogous to the von
              Neumann machine. We control the system by
              commanding an entity such as the UAV to go from
              the airfield to the oil spill, or to disconnect
              from one ground station to another ground station,
              to represent logically the transfer of control of
              the asset from one jurisdiction to another (in one
              of our exercises the the ground station was
              onboard of a Navy vessel). These controls are
              formally Bigraph Reaction Rules (BRR). Thus the
              oil spill monitoring exercise is abstractly a
              sequence of pictures or Bigraphs, evolving in
              response to the controls or environmental
              disturbances like network disconnections due to
              limited communication range. The âcyberâ
              components controlling the system are BigActors.
              These are mathematically Actors as defined by Agha
              and Hewitt bridged to the underlying Bigraph as
              defined in. BigActors are actors equipped with the
              ability to observe and control the Bigraph. Figure
              2(a) is an image captured from the UAV of the
              naval vessel and the emulated oil spill. The
              Portuguese Navy emulated the oil spill by
              releasing 100 kg of popcorn in the ocean (the
              yellow patch). The stills have been extracted from
              the UAV video to verify that the BigActor
              implementation is correctly able to control the
              UAV asset through take-off, three ground station
              handoffs, arrival in the oil spill area, imaging
              of the spill, connect with the AIS drifters,
              acquire their data, and deliver it back to base.
              The creative engineering contributions in the
              implementation of the BigActor model are an
              approximation of the BigActor semantics enabling
              distributed control, control by logical space
              programming, as opposed to physical space
              programming, and a process to build networks out
              of different vehicles built by different teams.
              Unlike the von Neumann machine, the structure of
              our ubiquitous computing machine changes as the
              oil spill exercise executes. These changes of
              structure are first locally observed by each
              vehicle in the system. These local observations
              are then propagated throughout the overall system,
              flooding the information in a distributed manned.
              Our BigActor implementation has a distributed
              Bigraph synthesis protocol and se- mantics to
              enable coordinated control with decentralized
              information. The implementation supports control
              by logical space programming because the
              distributed Bigraph synthesis pro- tocol provides
              each BigActor with a Bigraph abstraction of the
              whole system, enabling the actor to control at the
              Bigraph level. The Bigraph is a logical, not
              physical, space abstraction of the real world.
              Protocols in the implementation handle the binding
              of logical space to phys- ical space, ensuring
              Bigraph Reaction Rules actually have a physical
              effect. This is performed using the Robot
              Operating System (ROS) as the underlying
              middleware. We have created a ROS abstraction of
              each robot in the network called the ros vehicle
              and defined a set of libraries that make a robot
              into a ros vehicle. To put a new vehicle into the
              robot networks one should implement or install
              these libraries. They make the new vehicle into a
              ros vehicle, and thereby a node in the network.
              The network control is at the BigActor level. This
              is implemented using the Scala Actor Library
              running over ROS using ROSJava. The overall
              systemâs architecture is presented in Figure
              2(b).},
    URL = {http://terraswarm.org/pubs/123.html}
}

Posted by Christopher Brooks on 29 Sep 2013.

Notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright.