You are here

System of Systems (SoS) Certification Techniques


OBJECTIVE: To develop analysis techniques for the safety verification and validation of complex, autonomous system of systems (SoS) in order to prevent unintended interactions by addressing SoS challenges: interoperability, emergence, evolution, and non-determinism. DESCRIPTION: Safety-critical software being developed for collaborative autonomous aerospace systems is rapidly growing in size and complexity. Current methods of verification and validation (V & V) of this type of software lead to increased development costs, are time-prohibitive in testing, and create significant impacts to warfighter deployment and sustainment. This need is highlighted in the Air Force"s 2010 Technology Horizons,"It is possible to develop systems having high levels of autonomy, but it is the lack of V & V methods that prevents all but relatively low levels of autonomy from being certified for use."As today"s systems become more complex, there is a growing issue of unintended interactions within a single system and on a macro-level within a SoS. As systems are composed into a larger system, behaviors begin to emerge that did not exist at the individual or loosely coupled level. Therefore, one can easily see how the whole SoS architecture is greater than the sum of the parts. As the complexity of these more advanced systems increases, their non-linearity and non-deterministic qualities increase as well. This increased complexity can lead to instances of unintended interactions which may violate the safety and certification constraints of the system being developed. As systems become more tightly coupled, unintended interactions become more pronounced. An easy-to-understand example of this is the national highway system. Under light traffic, minor braking and acceleration among the various vehicles go largely unnoticed. As traffic increases, this braking and acceleration has a traffic wave effect through the entire system. This emergent behavior can be seen throughout most other SoS, such as air traffic control, internet traffic, aircraft development, operation of remotely piloted autonomous aircraft (RPA), automotive system development, and data-centric systems. A parallel example to the national highway system mentioned above is that of autonomous systems operating in the terminal area of an airport. As the dynamic, integrated, and rapidly changing environment of terminal area traffic fluctuates, the SoS interactions can lead to unpredictable results from autonomous systems. This lack of predictability (and non-determinism) causes a lack of trust and difficulty in certifying these systems. The overall vision for this research area is to reduce reliance on testing and enable SoS certification through trusted, formalized, and safe interactions of certified systems with focus both on single systems and within a SoS. As a step toward that goal, this SBIR topic looks to develop analysis and modeling techniques that focus on analysis of SoS versus exhaustive tests through technologies such as formal specification, boundary certificates, and predictions of system interactions. Through the use of the modeling tools and analysis techniques developed under this SBIR, it is suggested that more stable and easily scalable systems can be developed. The goal of this SBIR is not to develop a better low-level interface control document (ICD) management system, but rather focus on improved, system engineering modeling tools and analysis techniques that lead to formally provable and guaranteed safe interactions. Through the elimination of unintended interactions, especially during the evolutionary stage of a SoS life cycle, more trusted and autonomous systems will become a reality. PHASE I: Demonstrate the feasibility of a methodology and associated tools to enable SoS interaction analysis and future certification methods for autonomous systems. A description of the proposed system and associated methodology is required. Modeling and simulation may be needed to fully demonstrate system feasibility and operation. PHASE II: Continue to develop the tools and methodology proposed during the Phase I system. A challenge problem will be determined in the area of autonomous systems and a final Phase II demonstration will be conducted to inform interested personnel of the technology development in the area. PHASE III: Military Application: This technology has the ability to enable many advanced capabilities for unmanned aerial vehicles (UAVs), such as automated aerial refueling (AAR) and adaptive control, by providing safety guarantees for systems that cannot be exhaustively tested. REFERENCES: 1. Selberg, S., and Austin, M.A.,"Toward An Evolutionary System Of Systems Architecture,"Institute for Systems Research, University of Maryland. The International Council on Systems Engineering (INCOSE), 2008. Web. May 9, 2012.. 2. Technology Horizons - A Vision for Air Force Science & Technology During 2010-2030
US Flag An Official Website of the United States Government