Investigation of a Formal Fidelity Improvement Methodology for Rapid Re-Definition Simulation Systems, 07-R9699

Printer Friendly Version

Principal Investigator
Thomas Glass

Inclusive Dates:  04/01/07 – 04/01/08

Background - SwRI has developed a significant simulation methodology for aggregate modeling and simulation using rapid re-definition (RRD) technology, which has been applied successfully in multiple SwRI-developed applications including the Modeling of Aggregates of Individuals and Crowd Evaluation (MAICE) Station and Networked Sensors simulation. These applications allow users to rapidly develop and modify highly detailed scenarios, model complex behaviors, execute custom simulations, and perform analysis. This research and development effort seeks to define a formal methodology for improving simulation fidelity and validating scenarios for RRD simulations through a combination of automation and user-defined input.

Approach - RRD-based systems allow a user to iteratively examine scenario performance and rapidly make adjustments to scenario variables and entity behaviors as needed to improve the fidelity and realism of the simulation. These highly integrated systems have been designed and structured so that critical simulation control variables, scenario parameters, and entity behaviors are accessible and adjustable through user interfaces and data files. To date, the iterative, fidelity refinement process has been performed manually, a time-consuming and tedious process. The purpose of this research is to investigate, develop, and test a structured, organized methodology that identifies those aspects of the RRD process that can be automated and more clearly defines the role of the simulation developer in the fidelity improvement process.

Most commercially available simulation platforms cannot easily support scenario model re-definition or modification of modeled attributes; nor do they have the infrastructure and behavior model design required to build in such a capability. SwRI has developed the RRD technology to rapidly create and modify scenarios that have the potential to change the way military and commercial video games and simulations are developed. Using an existing SwRI-developed RRD modeling platform, this research will establish a formal methodology for automated fidelity improvement and validation of scenarios for RRD simulation systems.

Accomplishments - The project is being conducted in three phases. Phase I, which was recently completed, emphasized development of a basic methodology and identification of required mathematical and statistical components. Phase II tasks will include creation of a basic user interface that will allow a user to control the automated part of the methodology. In Phase III, the methodology will be finalized based on the results of a number of experiments. This will include several blind experiments and a final experiment comparing the results of the validation with results produced by the methodology generated by this research.

In the first step (1) of the Rapid Re-Definition Simulation System, clients and subject matter experts analyze and define real world data or events and provide information and metrics to build a simulation. After an initial scenario setup (2), modelers rapidly test (3), evaluate (4), and re-define (5) scenario components in order to improve accuracy and realism. This process is performed iteratively until the simulation scenario meets an acceptable level of fidelity (6).

2007 Program Home