Diagnostic Challenges Boeing's product lines span commercial and military aerospace vehicles, weapon systems, and systems of systems that are used around the world, by many diverse organizations, in a variety of harsh environments. Each of these Boeing products requires an optimized combination of autonomous embedded diagnostic capabilities to support real time operational decisions and near real time off board diagnostic capabilities for effectively driving maintenance decisions. The specific technical diagnostic problems that are encountered on individual programs are too numerous to discuss in this panel format. Rather, Boeing will describe the systemic process oriented diagnostic problems that are common across many platforms, and that can potentially be addressed by technologies developed in industry and academia. Representative problem areas include inefficient and inconsistent analytical processes, gaps between these analytical processes and the diagnostic design processes, and the lack of an effective process for maturing the diagnostic implementation over the product life cycle. (slides)
Bio Stan Ofsthun is a Technical Fellow in the Boeing Phantom Works organization. His career has primarily focused on the practical application of Testability, Built In Test, Integrated Diagnostics, and Integrated Vehicle Health Management (IVHM) technologies, processes, and tools to a variety of Boeing platforms, including F-15, F/A-18, X-45C and various space/weapons programs. He has also developed methodologies to integrate these efforts with reliability and safety disciplines, and is currently serving as the St Louis site technical representative on Boeing’s Reliability, Maintainability, and Systems Health (RM&SH) steering team. His educational background includes a B.S. in Electrical Engineering (Rensselaer Polytechnic Institute), a M.S. in Electrical Engineering (Washington University), and graduate certificates in Artificial Intelligence and Web Content Design (Washington University).
Diagnostic Challenges In the last decades, the fields of a telecom operator like France Telecom have changed: the main objective is not only to deploy and maintain a telecom network but to propose and to guarantee more and more innovative telecom services. The "Diagnosis" team mainly contributes to satisfy end-customers in two areas: the service supervision and the customer hotline.
- Services supervision: the objective is to collect and synthesize a huge amount of information coming from telecoms equipment (routers, probes...) in order to make them manageable by human operators. Purposes of this supervision could be QoS reporting, intrusion detection, fault detection...The main difficulties raise from the high number of sensors and the inherent distributed architecture but also from the knowledge acquisition since the system evolves continuously. Our approach based on chronicle recognition will also be evoked.
- Customer hotline: the objective is to give a diagnosis tool which can be used by customer and/or by technical support in order to satisfy as soon as possible the customer. One hard point is to give an "intuitive" modeling in order to be able to update the system with new services and/or problems but it is also required to explain the diagnosis in order to convince the end user. As this challenge is more recent for us, just some clues on what could be done to achieve that will be given to start panel discussion. (slides)
Bio After preparing his PhD at LAAS/CNRS (Toulouse) and graduated of the Ecole Polytechnique, Dr. Christophe Dousson joined France Telecom R&D in 1994 on a research position in Artificial Intelligence applied to Diagnosis. He was firstly involved in the field of fault management of telecommunications networks. Nowadays, the applicative fields are more related to virus and intrusion detection but the main topic of his work remains temporal reasoning for supervision and situations recognition. He also addressed the problem of automatic or assisted chronicle discovering based on log analysis and also based on a behavioral model of the system. Since 2004, he is also the head of the "Diagnosis" research team.
Diagnostic Challenges NASA designs, develops, and operates unique, highly customized spacecraft in uncharted territories. These spacecraft often experience failures due to operations outside design margins, extended exposure to extreme environments, and unforeseen interactions between the spacecraft and the external environment. Space operations provide a rich application domain for diagnostic technologies. However, there are several challenges that limit the applicability of cutting-edge diagnostic tools to space operations. These include: 1) verification and validation hurdles; 2) mismatch between computational requirements of advanced diagnostic algorithms versus the capabilities of space-qualified computing platforms; 3) lack of autonomous operations concepts for optimal decision making under uncertainty; and 4) requirements for extremely low false-positive and false-negative detection rates (especially in human space flight applications). (slides)
Bio Dr. Serdar Uckun is the Technical Area Lead for Discovery and Systems Health in the Intelligent Systems Division at NASA Ames Research Center. The technical area consists of approximately 70 researchers and engineers and focuses on engineering and scientific data understanding problems pertinent to NASA, including Integrated Systems Health management (ISHM). The technical area has the single largest ISHM R&D and systems engineering team at NASA, with over 50 people involved in ISHM-related tasks funded by NASA’s Exploration Systems, Aeronautics, and Space Operations Mission Directorates. In addition to his line management role, Dr. Uckun has been serving as the Project Manager for the ISHM project under NASA’s Exploration Technology Development Program. He participates in various NASA planning and strategy activities as a subject matter expert on health management. He has an M.D. from Ege University in Izmir, Turkey, M.S. in Biomedical Engineering from Bogazici University in Istanbul, Turkey, and a Ph.D. in Biomedical Engineering from Vanderbilt University in Nashville, TN. His technical interests include monitoring, diagnosis, prognostics, and scheduling. He has over thirty publications in peer-reviewed journals and conferences.
Diagnostic Challenges Automotive electronics and software represent a continuously increasing share in the added value of automotive products. Up to 90% of all automotive innovations are related to electronics systems. Increasing demands in vehicle safety, driver assistance, and comfort drives this trend. In addition legal requirements or environmental needs can only be fulfilled by using electronic hard- and software extensively. All vehicle domains are affected. As a consequence, the complexity of automotive electronics/system diagnosis is growing exponentially. The challenges also come from shortened development cycles, market competition and increasing customer expectation. We may have to reconsider the way of diagnosis.
Bio Dr. Liu Qiao is General Manager and Chief Technologist of Technical Research Dept. He is responsible for electrical, electronics and vehicle control related research at Toyota Technical Center, Toyota Motor Engineering and Manufacturing North America. Switching from a university faculty position, Dr. Qiao started his automotive career as an advanced automotive control system expert, advanced technology manager, eBusiness manager and research manager. Dr. Qiao successfully led Toyota's Canadian hybrid vehicle project and its market introduction. Dr. Qiao received Electrical Engineering Ph.D. from Tohoku University in Japan. He is an active member/supporter of many academic associations.
Panel ModeratorBio Johan de Kleer is a Principal Scientist in the Embedded Reasoning Area in PARC's Intelligent Systems Laboratory. His core interest is building a system which can reason about the physical world as well as he can. Until recently, he was Laboratory Manager of PARC's Systems and Practices Laboratory of Xerox's Palo Alto Research Center. This interdisciplinary laboratory conducted research ranging from social science to robotics. Two foci of the laboratory were: (1) Smart Matter - which exploits trends in miniaturization and integration to create a new generation of products and processes that benefit from the coupling of computational and physical worlds, and (2) Knowledge - knowledge management, which includes social science research on organizations, knowledge representation and understanding images and video streams. Johan received his Ph.D. from M.I.T. in 1979 in Artificial Intelligence. He has published widely on Qualitative Physics, Model-Based Reasoning, Truth Maintenance Systems, and Knowledge Representation. He has co-authored three books: Readings in Qualitative Physics, Readings in Model-Based Diagnosis, Building Problem Solvers. In 1987 he received the prestigious Computers and Thought Award at the International Joint Conference on Artificial Intelligence. He is a fellow of the American Association of Artificial Intelligence and the Association of Computing Machinery.