Integrating Automated Analysis into Development Practice
The Advanced Technology Center at Rockwell Collins has produced several formal methods results which have been key to high-assurance product success. We performed automated analysis for an Adaptive Display and Guidance system for commercial aircraft, analyzed the Green Hills INTEGRITY-178B real-time operating system kernel as part of an EAL6+ Common Criteria certification, and formally verified the AAMP7G microprocessor's built-in partitioning at the EAL7 level for certified secure systems. These are examples of design for verification, where the verification effort was aided by the fact that the development engineers obeyed certain design rules, utilized "safe" language subsets, documented their requirements and designs thoroughly, and so on. As formal methods continue to mature, industry leaders are moving beyond design for verification and toward design with verification. This approach integrates formal verification throughout the design process, thus identifying errors earlier, and avoiding costly rework. We will present current examples of high-assurance embedded system development at Rockwell Collins where formal verification is an integral part of the development process rather than after-the-fact analysis, and demonstrate the effectiveness of this approach. We will also discuss how our research partnership with Kansas State is helping us to achieve this vision.
David S. Hardin is a Technical Director in the Advanced Technology Center of Rockwell Collins. Dr. Hardin's published work includes papers in microprocessor architecture, formal verification, high-assurance systems, signal processing, as well as embedded and real-time Java. Dr. Hardin was a Rockwell Engineer of the Year for 1997, is the editor of the upcoming book Design and Verification of Microprocessor Systems for High-Assurance Applications (Springer, 2010), and holds 6 U.S. patents. In addition to nearly 20 years with Rockwell Collins, Dr. Hardin also served as Chief Technical Officer of aJile Systems for four years. He earned a Ph.D. in Electrical and Computer Engineering from Kansas State University in 1989.
