Engineering topics

Knight, John C.

Publications and source records attributed to Knight, John C..

At least 19 records

NASA NTRS2017

Comprehensive Lifecycle for Assuring System Safety

CLASS is a novel approach to the enhancement of system safety in which the system safety case becomes the focus of safety engineering throughout the system lifecycle. CLASS also expands the role of the safety case across all phases of the system's lifetime, from concept formation to decommissioning. As CLASS has been developed, the concept has been generalized to a more comprehensive notion of assurance becoming the driving goal, where safety is an important special case. This report summarizes major aspects of CLASS and contains a bibliography of papers that provide additional details.

Knight, John C.↗

NASA NTRS2014

Neither Pollyanna nor Chicken Little: Thoughts on the Ethics of Automation

This paper has raised issues concerning the ethics of automation in aviation systems, and outlined ways of thinking about the issues that may help in ethical decision making. It is very easy to be carried along by technology and the Pollyanna view, but just because we can do something, doesn't mean we should - which is perhaps a little milder than the Chicken Little view. Both views have merits, and we would view ethical decisions as ones that more appropriately balance or reconcile these conflicting viewpoints. We have set out some of the background to the problems of automation in aviation systems, but are aware that there is much more that could be said (considering military UAS, for example). We hope, however, that the brief introduction provides a foundation for the ethical questions that we have set out. The underlying aim in proposing ESCs is to make understanding ethical issues easier so that ethically-informed decisions can be made. Whilst we have not linked the discussion directly back to specific ethical decisions, we believe that making explicit those issues on which such judgments are based is a contribution to ethically informed decision making. We also believe that the four principles set out by the RAEng are reflected in this approach. We acknowledge that what we have set out, especially the ideas of ESC, goes some way beyond current practice and principles and there are significant technical issues to resolve before such an approach could be implemented. It is hoped, however, that the ideas will help improve the production and presentation of safety cases in a range of industries not just aviation - a Pollyanna view, of course!

Holloway, C. Michael↗

NASA NTRS2006

A Taxonomy of Fallacies in System Safety Arguments

Safety cases are gaining acceptance as assurance vehicles for safety-related systems. A safety case documents the evidence and argument that a system is safe to operate; however, logical fallacies in the underlying argument may undermine a system s safety claims. Removing these fallacies is essential to reduce the risk of safety-related system failure. We present a taxonomy of common fallacies in safety arguments that is intended to assist safety professionals in avoiding and detecting fallacious reasoning in the arguments they develop and review. The taxonomy derives from a survey of general argument fallacies and a separate survey of fallacies in real-world safety arguments. Our taxonomy is specific to safety argumentation, and it is targeted at professionals who work with safety arguments but may lack formal training in logic or argumentation. We discuss the rationale for the selection and categorization of fallacies in the taxonomy. In addition to its applications to the development and review of safety cases, our taxonomy could also support the analysis of system failures and promote the development of more robust safety case patterns.

Greenwell, William S.↗

NASA NTRS2005

Advanced Tools and Techniques for Formal Techniques in Aerospace Systems

This is the final technical report for grant number NAG-1-02101. The title of this grant was "Advanced Tools and Techniques for Formal Techniques In Aerospace Systems". The principal investigator on this grant was Dr. John C. Knight of the Computer Science Department, University of Virginia, Charlottesville, Virginia 22904-4740. This report summarizes activities under the grant during the period 7/01/2002 to 9/30/2004. This report is organized as follows. In section 2, the technical background of the grant is summarized. Section 3 lists accomplishments and section 4 lists students funded under the grant. In section 5, we present a list of presentations given at various academic and research institutions about the research conducted. Finally, a list of publications generated under this grant is included in section 6.

Knight, John C.↗

NASA NTRS2003

Risk-based Classification of Incidents

As the penetration of software into safety-critical systems progresses, accidents and incidents involving software will inevitably become more frequent. Identifying lessons from these occurrences and applying them to existing and future systems is essential if recurrences are to be prevented. Unfortunately, investigative agencies do not have the resources to fully investigate every incident under their jurisdictions and domains of expertise and thus must prioritize certain occurrences when allocating investigative resources. In the aviation community, most investigative agencies prioritize occurrences based on the severity of their associated losses, allocating more resources to accidents resulting in injury to passengers or extensive aircraft damage. We argue that this scheme is inappropriate because it undervalues incidents whose recurrence could have a high potential for loss while overvaluing fairly straightforward accidents involving accepted risks. We then suggest a new strategy for prioritizing occurrences based on the risk arising from incident recurrence.

Greenwell, William S.↗

NASA NTRS1999

Streamlining Software Aspects of Certification: Report on the SSAC Survey

The aviation system now depends on information technology more than ever before to ensure safety and efficiency. To address concerns about the efficacy of software aspects of the certification process, the Federal Aviation Administration (FAA) began the Streamlining Software Aspects of Certification (SSAC) program. The SSAC technical team was commissioned to gather data, analyze results, and propose recommendations to maximize efficiency and minimize cost and delay, without compromising safety. The technical team conducted two public workshops to identify and prioritize software approval issues, and conducted a survey to validate the most urgent of those issues. The SSAC survey, containing over two hundred questions about the FAA's software approval process, reached over four hundred industry software developers, aircraft manufacturers, and FAA designated engineering representatives. Three hundred people responded. This report presents the SSAC program rationale, survey process, preliminary findings, and recommendations.

Hayhurst, Kelly J.↗

NASA NTRS1998

Streamlining Software Aspects of Certification: Technical Team Report on the First Industry Workshop

To address concerns about time and expense associated with software aspects of certification, the Federal Aviation Administration (FAA) began the Streamlining Software Aspects of Certification (SSAC) program. As part of this program, a Technical Team was established to determine whether the cost and time associated with certifying aircraft can be reduced while maintaining or improving safety, with the intent of impacting the FAA's Flight 2000 program. The Technical Team conducted a workshop to gain a better understanding of the major concerns in industry about software cost and schedule. Over 120 people attended the workshop, including representatives from the FAA,commercial transport and general aviation aircraft manufacturers and suppliers, and procurers and developers of non-airborne systems; and, more than 200 issues about software aspects of certification were recorded. This paper provides an overview of the SSAC program, motivation for the workshop, details of the workshop activities and outcomes, and recommendations for follow-on work.

Hayhurst, Kelly J.↗

NASA NTRS1997

Why are Formal Methods Not Used More Widely?

Despite extensive development over many years and significant demonstrated benefits, formal methods remain poorly accepted by industrial practitioners. Many reasons have been suggested for this situation such as a claim that they extent the development cycle, that they require difficult mathematics, that inadequate tools exist, and that they are incompatible with other software packages. There is little empirical evidence that any of these reasons is valid. The research presented here addresses the question of why formal methods are not used more widely. The approach used was to develop a formal specification for a safety-critical application using several specification notations and assess the results in a comprehensive evaluation framework. The results of the experiment suggests that there remain many impediments to the routine use of formal methods.

Knight, John C.↗

NASA NTRS1995

Safety analysis

We are engaged in a research program in safety-critical computing that is based on two case studies. We use these case studies to provide application-specific details of the various research issues, and as targets for evaluation of research ideas. The first case study is the Magnetic Stereotaxis System (MSS), an investigational device for performing human neurosurgery being developed in a joint effort between the Department of Physics at the University of Virginia and the Department of Neurosurgery at the University of Iowa. The system operates by manipulating a small permanent magnet (known as a 'seed') within the brain using an externally applied magnetic field. By varying the magnitude and gradient of the external magnetic field, the seed can be moved along a non-linear path and positioned at a site requiring therapy, e.g., a tumor. The magnetic field required for movement through brain tissue is extremely high, and is generated by a set of six superconducting magnets located in a housing surrounding the patient's head. The system uses two X-ray cameras positioned at right angles to detect in real time the locations of the seed and of X-ray opaque markers affixed to the patient's skull. the X-ray images are used to locate the objects of interest in a canonical frame of reference. the second case study is the University of Virginia Research Nuclear Reactor (UVAR). It is a 2 MW thermal, concrete-walled pool reactor. The system operates using 20 to 25 plate-type fuel assemblies placed on a rectangular grid plate. There are three scramable safety rods, and one non-scramable regulating rod that can be put in automatic mode. It was originally constructed in 1959 as a 1 MW system, and it was upgraded to 2 MW in 1973. Though only a research reactor rather than a power reactor, the issues raised are significant and can be related to the problems faced by full-scale reactor systems.

Knight, John C.↗

NASA NTRS1994

Generalized implementation of software safety policies

As part of a research program in the engineering of software for safety-critical systems, we are performing two case studies. The first case study, which is well underway, is a safety-critical medical application. The second, which is just starting, is a digital control system for a nuclear research reactor. Our goal is to use these case studies to permit us to obtain a better understanding of the issues facing developers of safety-critical systems, and to provide a vehicle for the assessment of research ideas. The case studies are not based on the analysis of existing software development by others. Instead, we are attempting to create software for new and novel systems in a process that ultimately will involve all phases of the software lifecycle. In this abstract, we summarize our results to date in a small part of this project, namely the determination and classification of policies related to software safety that must be enforced to ensure safe operation. We hypothesize that this classification will permit a general approach to the implementation of a policy enforcement mechanism.

Knight, John C.↗

NASA NTRS1993

Development of a software safety process and a case study of its use

The goal of this research is to continue the development of a comprehensive approach to software safety and to evaluate the approach with a case study. The case study is a major part of the project, and it involves the analysis of a specific safety-critical system from the medical equipment domain. The particular application being used was selected because of the availability of a suitable candidate system. We consider the results to be generally applicable and in no way particularly limited by the domain. The research is concentrating on issues raised by the specification and verification phases of the software lifecycle since they are central to our previously-developed rigorous definitions of software safety. The theoretical research is based on our framework of definitions for software safety. In the area of specification, the main topics being investigated are the development of techniques for building system fault trees that correctly incorporate software issues and the development of rigorous techniques for the preparation of software safety specifications. The research results are documented. Another area of theoretical investigation is the development of verification methods tailored to the characteristics of safety requirements. Verification of the correct implementation of the safety specification is central to the goal of establishing safe software. The empirical component of this research is focusing on a case study in order to provide detailed characterizations of the issues as they appear in practice, and to provide a testbed for the evaluation of various existing and new theoretical results, tools, and techniques. The Magnetic Stereotaxis System is summarized.

Knight, John C.↗

NASA NTRS1992

Software development for safety-critical medical applications

There are many computer-based medical applications in which safety and not reliability is the overriding concern. Reduced, altered, or no functionality of such systems is acceptable as long as no harm is done. A precise, formal definition of what software safety means is essential, however, before any attempt can be made to achieve it. Without this definition, it is not possible to determine whether a specific software entity is safe. A set of definitions pertaining to software safety will be presented and a case study involving an experimental medical device will be described. Some new techniques aimed at improving software safety will also be discussed.

Knight, John C.↗

NASA NTRS1992

Development and validation of techniques for improving software dependability

A collection of document abstracts are presented on the topic of improving software dependability through NASA grant NAG-1-1123. Specific topics include: modeling of error detection; software inspection; test cases; Magnetic Stereotaxis System safety specifications and fault trees; and injection of synthetic faults into software.

Knight, John C.↗

NASA NTRS1991

Software fault tolerance using data diversity

Research on data diversity is discussed. Data diversity relies on a different form of redundancy from existing approaches to software fault tolerance and is substantially less expensive to implement. Data diversity can also be applied to software testing and greatly facilitates the automation of testing. Up to now it has been explored both theoretically and in a pilot study, and has been shown to be a promising technique. The effectiveness of data diversity as an error detection mechanism and the application of data diversity to differential equation solvers are discussed.

Knight, John C.↗

NASA NTRS1991

An experimental evaluation of software redundancy as a strategy for improving reliability

The strategy of using multiple versions of independently developed software as a means to tolerate residual software design faults is suggested by the success of hardware redundancy for tolerating hardware failires. Although, as generally accepted, the independence of hardware failures resulting from physical wearout can lead to substantial increases in reliability for redundant hardware structures, a similar conclusion is not immediate for software. The degree to which design faults are manifested as independent failures determines the effectiveness of redundancy as a method for improving software reliability. Interest in multi-version software centers on whether it provides an adequate measure of increased reliability to warrant its use in critical applications. The effectiveness of multi-version software is studied by comparing estimates of the failure probabilities of these systems with the failure probabilities of single versions. The estimates are obtained under a model of dependent failures and compared with the estimates obtained when failures are assumed to be independent. The experimental results are based on twenty versions of an aerospace application developed and certified by sixty programmers from four universities. Descriptions of the application, development and certifications processes, and operational evaluation are given together with an analysis of the twenty versions.

Eckhardt, Dave E.↗

NASA NTRS1990

An experimental evaluation of software redundancy as a strategy for improving reliability

The strategy of using multiple versions of independently developed software as a means to tolerate residual software design faults is suggested by the success of hardware redundancy for tolerating hardware failures. Although, as generally accepted, the independence of hardware failures resulting from physical wearout can lead to substantial increases in reliability for redundant hardware structures, a similar conclusion is not immediate for software. The degree to which design faults are manifested as independent failures determines the effectiveness of redundancy as a method for improving software reliability. Interest in multi-version software centers on whether it provides an adequate measure of increased reliability to warrant its use in critical applications. The effectiveness of multi-version software is studied by comparing estimates of the failure probabilities of these systems with the failure probabilities of single versions. The estimates are obtained under a model of dependent failures and compared with estimates obtained when failures are assumed to be independent. The experimental results are based on twenty versions of an aerospace application developed and certified by sixty programmers from four universities. Descriptions of the application, development and certification processes, and operational evaluation are given together with an analysis of the twenty versions.

Eckhardt, Dave E., Jr.↗

NASA NTRS1990

The use of self checks and voting in software error detection - An empirical study

The results of an empirical study of software error detection using self checks and N-version voting are presented. Working independently, each of 24 programmers first prepared a set of self checks using just the requirements specification of an aerospace application, and then each added self checks to an existing implementation of that specification. The modified programs were executed to measure the error-detection performance of the checks and to compare this with error detection using simple voting among multiple versions. The analysis of the checks revealed that there are great differences in the ability of individual programmers to design effective checks. It was found that some checks that might have been effective failed to detect an error because they were badly placed, and there were numerous instances of checks signaling nonexistent errors. In general, specification-based checks alone were not as effective as specification-based checks combined with code-based checks. Self checks made it possible to identify faults that had not been detected previously by voting 28 versions of the program over a million randomly generated inputs. This appeared to result from the fact that the self checks could examine the internal state of the executing program, whereas voting examines only final results of computations. If internal states had to be identical in N-version voting systems, then there would be no reason to write multiple versions.

Leveson, Nancy G.↗

NASA NTRS1990

Analysis of faults in an N-version software experiment

The authors have conducted a large-scale experiment in N-version programming. A total of 27 versions of a program were prepared independently from the same specification at two universities. The results of executing the versions revealed that the versions were individually extremely reliable but that the number of input cases in which more than one failed was substantially more than would be expected if they were statistically independent. After the versions had been executed, the failures of each version were examined and the associated faults located. The analysis showed that in some cases the programmers made equivalent logical errors, indicating that some parts of the problem were simply more difficult than others. The authors also found cases in which apparently different logical errors yielded faults that caused statistically correlated failures, indicating that there are special cases in the input space that present difficulty in various parts of the solution. A formal model is presented to explain this phenomenon. It appears that minor differences in the software development environment would not have a major impact in reducing the incidence of faults that cause correlated failures.

Brilliant, Susan S.↗