|
Original Date: 09/14/1998
Revision Date: 01/18/2007
Best Practice : Verification and Validation of Simulation Models
Raytheon Missile Systems Company (RMSC) developed a process that expedites software development by using the same software for modeling, algorithm development, simulation performance analysis, and flight software code development. The company also designed a process which uses a common model to validate software development, integration, and testing.
While developing flight software, RMSC discovered that most errors occur during the writing phase and typically create three problems: (1) failure to adequately verify the algorithms under realistic flight and operating conditions; (2) the hardware designer incorrectly communicates the defining algorithms to the software designer; and (3) failure to thoroughly test the software code at the unit, component, or configuration item levels. RMSC also noted that flight test performance predictions and post-flight performance analysis require a validation tool. Since flight tests are costly, the company strives to minimize the required number of tests and maximize the probability of success.
By developing algorithms in an accurate simulation and using the code directly in the flight software, RMSC prevented most errors from being incorporated into the software. This effort provided early identification of software failures, which significantly reduced costs associated with locating and correcting code errors later in the development process. RMSC also involves the analyst in the flight code development, which improves performance by addressing critical issues during the development of the in-real-time flight software. As a result, the company resolved many critical issues early in the development phase including data synchronization, throughput margins, data resolution, and interaction with other computer software units.
These software development guidelines enable RMSC to maximize the verification and validation of simulated models while reducing the number of tests required for performance evaluation. By following this approach, the company reduced the overall cost of code formulation, verification and validation, and final testing.
For more information see the
Point of Contact for this survey.
|
