NPS AUV Workbench Rehearsal Reality Replay for Unmanned
NPS AUV Workbench: Rehearsal, Reality, Replay for Unmanned Vehicle Operations Don Brutzman Naval Postgraduate School (NPS) Center for Autonomous Underwater Vehicle (AUV) Research Modeling, Virtual Environments & Simulation (MOVES) Institute 21 April 2008 1
Topics • • • Why modeling & simulation? AUV Workbench Components Sonar Visualization Technologies: X 3 D, XML, XMSF Looking ahead Demonstrations 2
theory = conceptual description of reality experiment = test theory in physical world Scientific method, 15 th-20 th centuries 3
model = formal representation of reality Scientific method, 1950 -present simulation = behavior of model over time Scientific method, 1950 -present 4
running together model = formal representation of reality Scientific method, 1950 -present simulation = behavior of model over time Scientific method, 1950 -present 5
Virtual environments can connect all models and simulations together Scientific method, emerging 21 st century Scientific emerging 21 st century 6
AUV Workbench Project Description • • Open source, Java, XML, X 3 D graphics Mission planning Robot mission execution Hydrodynamics response Sonar modeling 3 D visualization Compressed radio frequency (RF) and acoustic communications 7
AUV Workbench poster 8
Our 3 R’s: rehearsal, reality, replay • Same needs and capabilities for each: mission, visualization, data support, etc. • AUV workbench supports each § ongoing work, starting to mainstream • 13 years of effort is coming to fruition integrating great variety of successful work § new work projects occurring regularly § • Collaboration is welcome 9
Rehearsal Mission planning and preparation 10
Visualization, mission planning mission commands robot execution 6 DOF response hydrodynamics 11
Multiple vehicles supported 12
Rehearsal • Prepare missions, either manually or automatically via other software tools • Test robot software’s ability to perform commands • Test again with physics “in the loop” Hydrodynamics and control are critical, difficult § Sonar, environmental modeling § • Repeat until robust, with cautious respect § “Simulation is doomed to success” – G. Bekey 13
Mission views: iconic, tree, XML, dialog box Each view is consistent with GIS, 2 D, 3 D views 14
Supporting views: mission metadata, state 15
2 D planner: script missions 16
2 D planner: script missions Can edit missions by adding or removing script commands 17
2 D planner: agenda missions 18
Open. Map GIS display • Open. Map http: //www. openmap. org § § § Geographic Information System (GIS) Open source Java, bundled Building layers for areas of interest Geographic coordinates throughout Will synchronize with mission definitions, X 3 D 19
20
Secure sftp download of large GIS datasets 21
Reality: real-time mission support • Monitor mission progress • Task-level control using same mission vocabulary • Visualize and supervise operations § caveat, again: work in progress • Integrate acoustic and RF communications • Chat for distributed collaboration among participants, both human and robotic 22
Real-time mission data import/export • Export Mission commands that are already rehearsed § Convert to specific dialect particular to that robot § • Import Mission telemetry recording detailed track data § Data products: imagery, video, mission log, etc. § 23
Record mission metadata for archives • Support operator keeping detailed notes, kept in context when conducting mission • Prompt for full details as appropriate • Archive notes for later review and followup • Future work Automatic tests to confirm configuration, control § Automate pre-underway checklists § 24
Serial port communications • Configurable to different devices, ports 25
Java. Help support 26
Thesis list 27
Snapshot support 28
Sunspot • Sun Programmable Object Technology (SPOT) devices are small wireless Java microcontrollers that include accelerometers, USB port, LEDs and other connections • We are looking ahead towards using Sun. Spot devices as UUV, USV and UAV mobile robot controllers • http: //www. sunspotworld. com • https: //robotics. dev. java. net 29
Replay: post-mission support • Automatic archiving of mission to server § Being built into workbench – simplify user tasks • Integration and compression of all relevant data into single compressed XML file Metadata for mission § Many pieces: ordered mission, commands, telemetry, coefficients, contacts, etc. § Autonomous Vehicle Control Language (AVCL) is Ph. D. work by CDR Duane Davis § 30
Telemetry data replay 31
Geographic track plot 32
x y z versus t plot 33
phi theta psi versus t plot 34
XSBC compression of mission data • Compression of mission commands, telemetry • XML Schema-based Binary Compression (XSBC) • Take advantage of XML self-validation capability • Building composable sequence of filters for integrated data support 36
Automating server-side support Work is in progress… 37
ELumens dome display support www. elumens. com 38
Physical modeling • Control algorithms and 6 degree-of-freedom (6 DOF) hydrodynamics response • Sonar propagation, attenuation • Collision detection Direct vehicle contact and sensor contact § Separate use of same X 3 D graphics models § • Visualization greatly aids understanding § provides good “forcing function” for integration 39
Control algorithm coefficients 40
6 DOF dynamics coefficients 41
Wave modeling • Triple sinusoid Pierson Moskowitz equations provide good emulation of variable sea state Well understood example model § Many other variations exist, could substitute § • Real-time modeling of underwater vehicle response when broached Split hull into sections § Compute each one as linear approximation § Buoyancy components add to overall response § 42
REMUS mission search, from above 43
REMUS mission search, from behind 44
Group development support • • • Open standards throughout Open source Java for software All data structured as XML XMSF Bugtracker Email list with hypermail archive Online autoinstallers 45
XMSF Bugtracker 46
Mailing list support 47
Autoinstall support also linux and other unix 48
Environmental data inputs • Constant vectors for ocean current, wind • Net. CDF environmental data developed by NAVO/NRL Stennis supercomputer models • FNMOC web-services query to live/projected meteorological • Planned: add NITES/TED Services support 49
sonar-vis Project Description • Visualize multipath 3 D sonar propagation Situational awareness, sensitivity analysis § Multiple models: path, transmission loss, PD. . . § Operator familiarization, training, experience § • Enhance TDAs for at-sea operators Reachback using Web services messaging, accessing both computational and data assets § Open source open standards: Java, X 3 D, XML § 50
Sonar Visualization poster, I/ITSEC 2003 51
Goal Outcomes: sonar-vis project • Sonar and battlespace visualization • Link aircraft, ship and HPCC computing FNMOC, possibly Maui HPCC § Similar interfaces, with/without connectivity § “Tactical Supercomputing” refrigerator box § • Support Common Undersea Picture (CUP) Candidate technologies for spiral development § Have started usw-xml working group for USN § 52
Sonobuoy field visualization 53
Integrating 2 D/3 D interfaces with Web Services Participating in naval exercises 54
XML web services for METOC data 1 • Query panel and plotted response 55
XML web services for METOC data 2 • Monitoring initial query/response sequence 56
XML web services for METOC data 3 Server-side supercomputer response 57
XML web services for METOC data 4 58
Technologies • Extensible Markup Language (XML) Validatable data, binary compression § Web Services for message exchange § • Enhance current sonar-model engines Recursive Ray Acoustics (RRA) sonar computation § PC-IMAT/STAPLE/STDA(ASPECT) primary targets § Environmental data from FNMOC via Web services § • Extensible 3 D (X 3 D) Graphics § Open-standard open-source interactive visualization 61
XML in 10 Points http: //www. w 3. org/XML/1999/XML-in-10 -points • XML is new, but not that new • XML leads HTML to XHTML • XML is modular • XML is basis for RDF and the Semantic Web • XML is license-free, 400+ member companies & institutions platform-independent and in World Wide Web Consortium (W 3 C) well-supported already understand the business case • XML is for structuring data • XML looks a bit like HTML • XML is text, but isn't meant to be read • XML is verbose by design • XML is a family of technologies 62
Extensible Modeling & Simulation Framework (XMSF) • Web services for all manner of M&S • A composable set of standards, profiles, and recommended practices for web-based M&S • Foundational precepts: Internet network technologies, Extensible Markup Language (XML)-based languages, and service-oriented architectures for simple messaging • Enable a new generation of distributed M&S applications to emerge, develop, interoperate with tactical systems • Many easily repeatable exemplars using Web Services http: //www. Moves. Institute. org/xmsf 63
What is 3 D? • 2½D works for chart-oriented displays • 3 D gives “fly-thru” freedom of viewpoint View physically based propagation paths § View depth separation § View bottom, surface interactions § View multiple overlapping sensors § • Augment (not replace) existing displays 64
What is X 3 D? • Extensible 3 D (X 3 D) Graphics Virtual Reality Modeling Language (VRML) updated § Third-generation ISO specification § Compatible XML. x 3 d and Classic VRML. wrl encodings § • Deliverables § § § Specification updates, with compatible XML tagset Multiple implementations, including open-source Scene Access Interface (SAI) strongly typed API Conformance suite and examples Authoring capability: X 3 D-Edit, using XML for XML… 65
Further X 3 D motivations • Authoring is hard, “Content is King” § X 3 D is not competing with specialty formats, instead provide common interoperability/interchange § Strong validation checks eliminate most authoring errors before content escapes § Plays well with next-generation Web languages “ 3 D hardware problem” is already solved 66
X 3 D Specifications honeycomb diagram X 3 D Specification itself is componentized and extensible 67
X 3 D-Edit complete interface Context-sensitive, self-validating, multi-lingual editing tools 68
SAVAGE 3 D Model Archive Lots of models! 69
Numerous underwater vehicle models 70
Looking ahead: more technologies • XML Schema-based Binary Compression (XSBC) better than zip, adds validation § Adding Forward Error Correction (FEC) Current work • Tactical supercomputing • XML Tactical Chat (XTC) 71
XML Schema-based Binary Compression (XSBC) • XML encoding for validation benefits • XML schema contains adequate information to autogenerate comprehensive encodings • Tokenization of elements, attributes • Strong data typing of value payloads • Lossless • More efficient than compressed numeric text 72
XML Schema-based Binary Compression (XSBC) 3. 3 MB 16. 6 MB 38. 4 MB 11. 6 MB 2. 7 MB originals Conclusion: XSBC performance already better than zip! 73
XSBC comparison tool 74
Forward error correction (FEC) • Added redundancy allows receiver-side detection & correction of message errors Many military channels are noisy RF links § Avoids “retry until you die” on acoustic links § Big help on long-latency, low-bandwidth links! § • Hamming FEC is one technique of several § Re-exploring Stephen Reimers 1995 thesis “Towards Internet Protocol (IP) over Seawater” 75
Tactical supercomputing • Linux clusters can create new resources 5 off-the-shelf new PC boxes + disk storage § $10 K = 10 processors = 10 Gigaflop § Refrigerator-rack footprint easily fits shipboard § Industry can provide even higher capabilities § • Exploring intermediate-level resources for previously supercomputer-level problems § Consistent access via grid/web services 76
XML Tactical Chat (XTC), I/ITSEC 2003 poster 77
Chat log window 78
Chat log debug mode shows underlying XML 79
Event monitoring via instant messaging chatbot listens and reacts to free-form messages of interest by plotting mine onto chartlet 80
Java 1. 4. 2 regular expression parser on chat: Breakdown of regular expression pattern: Regular Expression ^. * Any characters or words before "MINE" (? i)MIN E [s|S] ? “s” or "S" occurs once or none at all Caseinsensitive word “MINE” . * [ (]{1, 2} 1 to 2 characters "space" or "(" Any character Group 1 (digit) (d* ) , [ ]{0, 2} Comma followed by 0 to 2 character "space" Group 2 digit) (d*) , [ ]{0, 2} Comma followed by 0 to 2 character "space" (d*) [ ). ]? + One or no character "space", fullstop or ")" Group 3 digit) Meaningful messages can be extracted from chat text, thus enabling automatic structure for user support 81
Related work: usw-xml • The usw-xml working group is improving Undersea Warfare (USW) interoperability using Extensible Markup Language (XML) tagsets for system data interchange. • Cooperative collaboration between many stakeholders is needed to achieve good interoperability. 82
Motivation • Many XML technical capabilities enable significantly improved capabilities for USW system interoperability § Connecting legacy systems, diverse partners § USW Decision Support System, other projects § • We expect this work to broadly benefit the Navy, industry and scientific community. 83
Upcoming: AUV Fest 2008 • Naval Undersea Warfare Center (NUWC) Newport RI, 12 -24 May 2008 • Sponsors: NOAA and ONR • 17 underwater robots and various surface support craft exploring archaeological sites • AUV Workbench supporting data collection, reporting, visualization 84
AUV Fest sediments data 85
AUV Fest 2008 86
AUV Fest 87
AUV Fest 88
AUV Fest NUWC Waterfront 89
Summary • Significant collected AUV capabilities § • • • Support rehearsal, reality, replay Integrated as tactical application Open standards: XMSF, X 3 D, chat, etc. Open source + commercial compatibility Improved messaging, net-centric exemplar We hope to add all possible vehicles! Collaboration and questions welcome 90
Acronyms • 3 D: Three dimensional • 6 DOF: Six degrees of freedom (x y z, roll pitch yaw) • AUV: Autonomous Underwater Vehicle • AVCL: Autonomous Vehicle Control Language • CD: Compact Disk • CUP: Common Undersea Picture • FEC: Forward Error Correction 1 • FNMOC: U. S. Navy Fleet Numerical Meteorological & Oceanographic Center • HPCC: High-Performance Computing Center • Java: programming language • METOC: meteorological and oceanographic (data) • NAVAIR: U. S. Naval Air Systems Command 91
Acronyms • NPS: Naval Postgraduate School, Monterey California • PD: Probability of detection • RF: radio frequency • RRA: Recursive Ray Acoustics Sonar Propagation • SBIR: Small Business Innovative Research • TDA: Tactical Decision Aid • USW: Undersea Warfare 2 • X 3 D: Extensible 3 D Graphics Specification • XML: Extensible Markup Language • XMSF: Extensible Modeling and Simulation Framework • XSBC: XML Schema-based Binary Compression • XTC: XML Tactical Chat 92
Contact Don Brutzman brutzman@nps. edu http: //web. nps. navy. mil/~brutzman Code USW/Br, Naval Postgraduate School Monterey California 93943 -5000 USA 1. 831. 656. 2149 voice 1. 831. 656. 7599 fax 93
- Slides: 90