Loading...
Please wait, while we are loading the content...
Similar Documents
Proven Techniques for Efficiently Generating and Testing Software
| Content Provider | Semantic Scholar |
|---|---|
| Author | Wegner, Keith R. |
| Abstract | The MATLAB programming language is quickly and unobtrusively becoming a primary prototype and analysis tool at Northrop Grumman Electronic Sensors & Systems Sector (ESSS) as well as at other engineering entities the world over. Its value to both systems and software engineers has been well demonstrated on many programs at ESSS, including Comanche, and it is especially well-suited to effectively support the corporate Integrated Product Team (IPT) concept. Engineers desiring to be technologically competent and efficient may soon find a solid foundation in MATLAB to be almost indispensable. Its application is becoming pervasive within systems engineering groups, and it is rapidly spreading to other disciplines, such as software, rapid pro-totyping, and financial analysis. On a similar note, MathCAD is routinely used in the systems engineering disciplines to perform detailed algorithm analysis and development. Although not as well-suited for large, elaborate simulations as MATLAB, this tool excels in performing symbolic manipulations and in developing and documenting detailed mathematical derivations. It is especially renowned for its representation of equations using succinct, precise mathematical notation. Systems engineers are routinely using both MathCAD and MATLAB to develop and capture their design documentation in Mode (MDD) and Function (FDD) Description Documents. Thus, software engineers with a working knowledge of MathCAD and MATLAB will be well positioned to much more efficiently develop software products by bridging the gap from systems MDDs and FDDs to embedded software. These two tools are especially powerful when applied to processes that are inherently mathematical, such as stochastic processes, image and signal processing, inertial navigation, etc. This paper describes informal techniques developed and successfully applied by the author, within the governing software process, on the Comanche program at ESSS to produce robust, optimal software directly from a systems FDD captured in a MATLAB simulation. In this process, MathCAD was first used to quickly generate the algebraic equivalent of involved Kalman filtering matrix equations expressed in MATLAB. These expanded representations were then coded and tested in MAT-LAB before being translated into the target Ada programming language using available text-editing tools. MATLAB was extensively used to rigorously unit test the deliverable Ada software product. This deceptively simple process allowed the author to design and develop large, efficient amounts of code in a very short time. Furthermore, the final software was found to be virtually error free, making the successful unit testing of code so developed almost a foregone conclusion. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.crosstalkonline.org/storage/issue-archives/2000/200006/200006-Wegner.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
