Next Telecon: Every other Wednesday 10am PT 1pm ET (Standing OSSE Status Meeting)

2009-08-12 Telecon (OSSE Status Meeting)

Participants: Oscar, David, Stephanie, Wayne, Michael

Discussion Topics:

• Michael has updated the OSSE architecture and the OSSE integration data flow pages on the Wiki. All OSSE contributors should review these materials for consistency and contribute any details for refinement
• Walk through the OSSE integration data flow. There is shared general understanding among the OSSE team members. A few open questions remain
• Oscar and Michael will talk with Yi about the integrated user experience of the virtual ocean
• Oscar and Michael will talk with Pierre and Waybe about the MIT MSEAS model integration
• UCSD still needs to request the JPL tool licenses
• CI Kick-Off meeting: the OSSE has scheduled time for a presentation and demo in the morning of Sept 10. The time slot will be about 45 minutes. It will cover 10 minutes OSSE design and integration (Oscar, Michael), 10 min demo model integration and virtual ocean (Yi), 10 min demo shore planning (David), 10 min demo dynamic simulation and advanced on-board AUV autonomy (Arjuna/Stephanie) and the rest for Q&A
• All please provide presentation materials to Michael and Oscar for integration
• Virtual simulation: Stepwise integration of OSSE components. Oscar and Michael prepare the timeline, starting with observational data product integration
• One of the drivers of the OSSE experiments is the availability of multiple alternative (potentially conflicting) data products for mission planning. The tools and user interfaces should reflect this and let the scientist/glider pilot explore and refine various alternatives, before deciding on one plan for glider tasking. Tools should be able to interface to various data products and keep track of multiple alternative sets of plans/parameters in parallel.
• Stephanie will evaluate with Wayne how to make use of various different forecast model products (MSEAS, HOPS, ROMS, POM)
• The Shore planning (Path + detailed planning in ASPEN) components use a rough ocean current model for initial planning. The precise planning is expected to come out of the MOOS-IvP driven dynamic simulation. CASPER is one process in the MOOS environment. The dynamic simulation environment (driven by MOOS, consisting CASPER) is a set of processes that can be started by the planning user iteratively as often as needed for refinement.
• Uncertainty in actual glider position when surfacing is addressed by developing conservative plans (e.g. using larger watch circles). Such plans should work in most of the cases; in some unexpected situations, new plans need to be developed during the surfacing interval.

2009-08-03 Telecon (OSSE Status Meeting)

Participants: Oscar, David, Stephanie, Wayne, Yi, Michael

Discussion Topics:

• Yi suggests a stepwise integration before the dry run. This way each group and technology gets specific attention
• Michael will prepare a flow diagram with the integration steps
• David would like to get exemplar data products that resemble the data products of the integrated OSSE
• For glider mission file generation, we can use the Rutgers docserver. It expects a list of waypoints and will automatically fill the glider inbox with a medium priority waypoint mission file. Operators have to ability to submit high priority overriding behaviors. Allows for graceful degradation.
• John Kerfoot is back from vacation. Oscar will arrange a telecon with him, David and Arjuna to discuss the stock-glider operations interface.
• Need to discuss the role of MOOS-IvP within the simulation where it can't directly influence glider control. Can it validate or even refine behaviors in a shore-side control loop?
• Target dates for the demos are: Virtual simulation (dry run) Sept 21-25; field deployment the week of Nov 7
• OSSE will prepare a demo for the OOI Kick-Off meeting Sept 9-11 in Denver.

2009-07-01 Telecon (OSSE Status Meeting)

Participants: Oscar, David, Michael

Discussion Topics:

• Oscar provided his input beforehand to Michael
  • Oscar has contacted a European group recommended by Webb Research. This group has apparently done some work on glider behavior models. No response yet
  • MIT's modeling group (Pierre/Wayne) is in contact with Oscar to get numerical model references. Oscar refers to numerical modelers
  • Oscar works on a data flow overview and per week time line, showing the incoming observational data and forecast data products within a typical week. Will document on Confluence next week. This should provide answers to when/which forecast models are run and provide a baseline for the experiments and scheduling.
  • The simulation will follow the same timeline. Historic data can be fed into the simulation at expected times according to the time line
  • Resource questions will be decided next week within the OOI CI management team.
  • The NY Stevens HOPS model made good progress recently with the help of Rich Signell to be compatible with the THREDDS data server. This data product is operational.
  • The UMass Dartmouth data product is available as well
  • Oscar suggests an OOI publication with authors from the OSSE group at EOS. It should show an end-to-end view on interactive ocean observing with a sensor network and could draw from previous publications by David (current paper) and Steve (SensorNet), as well as the OOI CI architecture, including the full Planning&Prosecution subsystem.
  • Oscar suggests the last week of September for the virtual simulation experiment.
  • Oscar suggests mid of November for the field experiment. There is an OOI Project Scientist meeting with greater community participation scheduled in this time frame and the OSSE field deployment could provide a live demo.
  • For the field deployment, other groups have announced their interest in participation. Ship time is available. We have lots of observational data and forecast data products available in several formats. We have 4 gliders for survey and 1 taskable stock glider, driven by the OSSE integration architecture.
• Status JPL (David)
  • David's questions of last week regarding time warp and glider behaviors could be resolved with Arjuna
  • Current AUV behaviors are not fundamentally different than expected glider behaviors. Can work with existing behavior set right now
  • Working on integration of tools for shore-side simulation, including Google Earth as described in the paper and on Confluence
  • Current plan scope about 24 hours. A more rough underlying week long observation plan is thinkable with similar characteristics if longer term goals are persued (e.g. transit over several days)
  • Open Issue: Who translates the shore-side plan into glider stock behavior commands and mission files? (e.g. waypoints might have to be set far away in order to avoid oscillation)
• Publication
  • David plans to submit a paper (Mission Planning in a Dynamic Ocean Sensorweb) for SPARK with OSSE group authors. Submission is end of this week. If the paper is accepted, revisions can be made for the camera ready version.
  • Michael will review this week and provide input
• Open Issues
  • UCSD delayed in requesting ASPEN/CASPER license. Will work on that when Matthew is back next week.
  • Possible interim solution: MIT as licensee could provide the license for the project. There are still some issues to be resolved between JPL and MIT before MIT license can be issued.

2009-06-17 Telecon (OSSE Status Meeting)

2009-06-03 Telecon (OSSE Status Meeting)

2009-05-20 Telecon (OSSE Status Meeting with OSSE Modeling Group)

Face-to-face Meeting: 2009-04-28/29 at Rutgers University

Agenda

• Day 1 (Tuesday, 4/28): CI Day
• Day 2 (Wednesday, 4/29): OSSE Modelers Day

2009-04-02 Telecon (OSSE Status Meeting)

Participants: David, Yi, Michael

Discussion Topics:

• Ocean forecast model integration with path and schedule planner
  • Yi has provided ocean model data from the SCCOOS region to Peggy and David. Data are forecast predictions for ocean currents as hypercube with km resolution on the surface.
  • Numericial model data has been cleared and is available from JPL
  • Data could be successfully integrated for path planning on shore. Strategy is to use numerical models combined with Matlab for path planning of Gliders. Path planning is modularized and separated from schedule planning.
  • The result is a sequence of waypoints as input for shore-side ASPEN path planner. ASPEN will refine the waypoint plans and augment with glider specific planning.
  • ASPEN output plans are compact in size, and can be transmitted to gliders as command messages during surfacing periods
• JPL focus shifted after successful proof of concept integration of CASPER with MOOS IvP Helm to path planning of glider and planning of glider resources.
• Specific requirements for the OSSE numerical model by Yi will be determined at the Rutgers meeting
• OSSE face-to-face meeting at Rutgers:
  • Financial organization will be done by WHOI
  • Several numerical modelers have accepted the invitation. Modelers will be well represented.
  • Additionally present will be Rich Signell (USGS, with NOAA) and Eoin Howlett (OPenDAP)
• Rutgers glider data and mission files availably on the OSSE wiki today; provided by John K
• Oscar is continuing to get the glider simulator to JPL asap

2009-03-19 Telecon (OSSE Status Meeting)

Participants: Arjuna, Stephanie, Wayne (MIT), Yi, Steve C, Peggy, David (JPL), Oscar, Michael M (CI)

Discussion Topics

• David has posted a video about the integration of CASPER and MOOS-IvP behaviors in a glider simulator environment on the OSSE Wiki
• The dates of the OSSE face-to-face meeting have been determined: Tuesday 4/28 (full day) for the CI part and Wednesday 4/29 (full day) for the modeling topics
• Glider simulator provided by Rutgers and others:
  • The first simulator is available to MIT. There is a delay in delivering the second but it should arrive soon at JPL.
  • Rutgers has years of experience with the simulator and can provide support.
  • The simulator consists of hardware and software. Main function is to take input from simulated instruments (velocity at different depths) and compute a flight path and interact with on-board processes, such as the IvP-Helm. Most interaction is with vehicle control systems. The simulator also provides a glider movement model through its software environment.
  • MIT has a comparable Remus AUV simulator that could be used if needed as well
  • For early integration, it is possible to run CASPER independent of the glider simulator just with a ROMS environment simulation (does not even have to be accurate). The integration CASPER with MOOS-IvP still does not require the glider simulator environment. Main interface is between MOOS-IvP and the glider simulator.
  • Arjuna has a glider emulator box if needed. Good to understand gliders and the possibility of adding a computation module running CASPER etc. Does not integrate with environment numerical model.
• Mike B will set up a SVN code repository at MIT serving as main code integration environment for all OSSE participants.
• Oscar will make available example glider flight data and engineering files (mission files). Will then set up a telecon with John Kerfoot, Rutgers systems engineer to clarify details.
• Participant status update
  • MIT (Arjuna): Glider simulator received, Stephanie works on glider specific behaviors
  • JPL (Steve/David): Work on CASPER integration (see video); held telecon with MIT to division labor between planner and MOOS-IvP; will hold meeting to discuss model integration with Yi.
  • JPL (Yi): Gets in touch with modelers about ROMS specifics and models to use; will clarify details during and after Rutgers meeting. There is a good amount of model work existing in the MA bight and OSSE does not want to repeat this effort. Rather draw from it and complement it (e.g. by running existing models in ensembles).
  • CI (Oscar): Invited several modelers to the Rutgers meeting. Run gliders continuously. Prepare face-to-face meeting.
  • CI (Michael): Following integration process and getting educated about component technologies; Request of JPL software still pending
• One success criteria CI for OSSE is the successful training an non expert team member at UCSD to work with IvP behaviors and CASPER models. Current target: Michael
• The CI has two relevant prototype efforts ongoing that will feed into the OSSE
  • The Messaging Service will provide the messaging infrastructure to connect component technologies. This service should also be deployed onboard, in a light-weight configuration
  • The Data Exchange will provide numerical model (and other) data upon request to consumers in various formats, such as NetCDF and DAP. It's the binding element between data producer and data consumer, providing integration with the rest of the CI.
  • The OSSE integration efforts should reflect the design of these CI integration components but not be dependent on the availability of any of the prototype implementations. CI will update on the design periodically and feed in software results when available. OSSE should go ahead with direct integration of components as needed.
• The date of the next telecon is targeted for 4/2. The time will be determined by Rita using Doodle.

Action Items

• (Oscar) Prepare Rutgers meeting
• (Oscar) Follow up on glider simulator; make sure JPL gets it asap. Send to Steve Chien.
• (Oscar) Make available regional glider flight data on the OSSE Wiki; also provide initial and updated mission files, command sequences etc.
• (Oscar) Schedule a telecon with John Kerfoot (Rutgers) about glider data and engineering
• (Mike B) Set up shared code repository for all OSSE participants
• (Matthew) Request JPL software
• (Michael/Rita) Determine date of next telecon

2009-03-16 Telecon (MOOS IvP Behaviors Telecon)

Participants: David, Arjuna, Stephanie Petillo, Mike B (partly), Michael M

Discussion Topics

• Stephanie is new in Arjuna's group at MIT, working on the development of OSSE specific IvP behaviors
• JPL has developed an initial CASPER-MOOSDB integration building on an existing MOOSDB module.
• Discussion and background on MOOS-IvP behaviors and structuring. Control messages and MOOSDB variables to control behaviors. Current structuring as tree of states, see IvP documentation.
• Feedback variables produced by the MOOS IvP-Helm to be used by CASPER (via MOOSDB variables) exist
• OOI CI integration infrastructure: On-shore and on-board messaging system, data distribution network. For both, Michael will provide specifications and details once they become available.
• Current existing behaviors are mostly AUV specific. Need adaption for glider capabilities and physics. Possible glider behaviors: Waypoint (~1 km range), yoyo. In the mid-term, yoyo based on sensor data processing to detect the mixed layer boundary.
• Possible planner control: Surfacing decision at waypoint, depth of vertical profile, turn on/off sensors
• Anonymous access is possible to get MOOS-IvP software tree from MIT SVN server
• (Mike B) MIT will set up a OOI OSSE specific file repository on their servers for all to share code files.
• The OSSE team should develop clear understanding of intended usage, scope and arrive at some ConOps. This will then guide the determination of behaviors and operational modes.

2009-03-05 Telecon (Kick-Off)

Participants: Arjuna, Mike B, Yi, Steve, Michael M, David, Oscar, Pierre

Discussion Topics

• Operational Model environment for MA bight. There are three models available for OSSE.
  • NY harbor model: extends on the shelf, focused on highly resolving features in the NY harbor, majority continental shelf
  • Rutgers ROMS: Covers Nantucket to gulf stream, Chesapeake Bay, NE snapshop
  • Dartmouth Model: Wiget; gulf of Maine down to MA bight)
• Model data exchange format: NetCDF and THREDDS catalog. Catalog available at http://tashtego.marine.rutgers.edu:8080/thredds/catalog.html . Yi will make data available using his database in formats to be determined. A NOAA data server is currently being set up.
• Glider simulator: One simulator sent to Arjuna, another to be sent to Steve once available. Rutgers can provide assistance with the glider systems and payload options. The platform is Linux
• MOOSDB processes: Processes communicating with glider devices (ACOMM, Satcomm) is the typical scenario. A repository of such processes exists, for instance for an Iridium modem.
• MOOS IvP behaviors: Details to be determined in another telecon. Parameterization through comm messages (via ACOMM, Iridium) to enable behaviors or through MOOSDB variables to switch behavior modes.
• Route planning: There are publications and software for glider/AUV route planning using hyphen/balloon trajectory tracking. Available from Pierre (MIT) and within JPL.
• Collaboration tools: The OSSE Wiki is established. The Jira tool can be used for task assignment and tracking. Code repositories will be set up for any source code. Michael M is the point of contact.
• Next meetings:
  • The OSSE face-to-face meeting will be held at Rutgers mid/end of April. One full day for the OSSE/CI part and one day for the modeling group (before or after). Michael and Rita will coordinate availability with all participants. Target date is the week of April 13. OSSE group participants should block that time for travel and meeting.
  • Next status update telcon: In two weeks from now. Michael and Rita will coordinate date and time.
• Physical deployment experiments: This fall (end of October) seems a good time frame. Oscar has ship time and possibilities for experiments in the MA bight.

Action Items

• Oscar: Provide overview of existing and planned operational models for the MA bight together with their scope, resolution and purpose
• Oscar: Evaluate existing datasets from the MARCOOS domain to be used by the JPL mission planner
• Arjuna, David: Hold telecom on IvP behaviors and their parameterization
• Mike B: Provide current documentation for MOOS-IvP behaviors and Helm
• Matthew: Request CASPER/ASPEN software from JPL and provide in project environment
• Oscar: Provide a list of atomic commands sent to gliders via ACOMM/Satellite
• Michael/Rita: Coordiate date/time for next status telecon and face-to-face meeting in April
• Michael: Set up email list for OSSE group and for OSSE modeling group

2009-01-15 Telecon (OSSE Preparation Call)

Discussion Topics:

• OSSE contract and statements of work
• Budget and planning for FY09 and FY10. Focus in FY09 is integration proof of concept in simulator environment. Field deployment planned for FY10.
• Core design and integration strategy