A Graphical Approach to Simulation Interoperability

1
A Graphical Approach to Simulation
Interoperability

• Christopher Cull
• Jorge Valencia

2004 Fall SIW
2
Agenda

• Graphical Programming vs. Graphical Simulation
  Environments
• Overview of CAE ROSE
• Rationale for a Graphical Approach to Simulation
  Interoperability
• A Brief History of DIS and HLA
• Developers vs. Modelers
• The Challenge
• and the Solution
• Conclusions Future Directions

3
Graphical Programming

• What is Graphical Programming?

A programming environment in which graphical or
iconic elements are manipulated by the user in an
interactive way according to some specific
spatial grammar for program construction.

• essentially, programming by drawing a diagram
  instead of writing text
• Also called visual programming or iconic
  programming

4
Graphical Simulation Environments

• Many graphical programming tools for general
  purpose programming, frequently using UML
• Simulation-specific tools include
• MATLAB/Simulink
• CAE ROSE
• Most are hybrid graphical programming
  environments
• Some components defined textually
• Models defined graphically
• Most have little or no support for HLA or DIS

5
Overview of CAE ROSE
6
Benefits of Graphical Simulation using CAE ROSE

• Reliable code
• Validated objects code generators
• Model details are programmer independent
• More intuitive models
• Easy to understand for new modelers on project
• Schematics closely resemble design drawings
• Cost Schedule reductions
• Code generators automatically create all required
  simulation components
• Code, run-time database, documentation
• Animated schematics

7
A Brief History of Simulation Interoperability

• Distributed Interactive Simulations (DIS)
• Low-level Protocol Data Unit (PDU) specification
• High Level Architecture (HLA)
• High-level Application Programming Interface
  (API) specification
• Sample lines of code for a small reference
  simulations interoperability needs
• Using DIS 10,000 lines (includes networking
  code)
• Using HLA 8,000 lines (but much thinner code)
• Using typical HLA framework 200-300 lines

8
The Hidden Cost of Less Code
9
Developers vs. Modelers

• Developers are programmers relevant skills and
  knowledge usually include
• In-depth knowledge of one or more programming
  languages such as C or Java
• Object-oriented design principles
• Modelers are engineers and subject matter experts
• Understand the system to be modeled
• Frequently little programming knowledge

10
The Great Divide

• In many traditional methodologies, models are
  specified by modelers, but implemented by
  developers
• Common problems
• Language imperfect translation of modelers
  knowledge into specification
• Optimal implementation constrained by
  specification due to lack of knowledge of
  programming implications
• Imperfect understanding of specification by
  developer due to lack of knowledge of system
  behavior
• Impact on schedule and budget

11
The Challenge

• Generic support for HLA in a Graphical Simulation
  Environment
• Zero lines of hand-written code
• Implementation by non-programmers (modelers)

12
The Solution

• Implemented an HLA Code Generator for CAE ROSE
• Automatically creates HLA interface code for a
  simulation created in ROSE
• Modelers add HLA interface objects to their
  simulations from pre-defined libraries

13
HLA Objects in ROSE

• A set of one or more ROSE libraries represents a
  particular FOM
• Simple library objects can be created by modelers
• Example a proxy object for an HLA object
  simulated in ROSE
• However, more complicated library objects
  probably need to be created by developers
• Some C coding still required, on par with using
  a typical HLA framework
• Example a proxy object representing multiple HLA
  objects simulated outside of ROSE

14
Example Graphical Implementation of an HLA
Interface

• Simple aircraft model

• Publish position and orientation only

• Note inputs to proxy object are from various
  parts of the ROSE model

15
Conclusions Future Direction

• Successfully implemented HLA within the Graphical
  Simulation Environment, CAE ROSE
• Further Work
• Complete support for all HLA services (ownership
  management, data distribution management, etc.)
• Continue to reduce amount of programming required
• e.g. import FOM to create libraries

16
(No Transcript)