Engineering Commons^™

Bayesian Software Health Management For Aircraft Guidance, Navigation, And Control, Johann M. Schumann, Timmy Mbaya, Ole J. Mengshoel

Ole J Mengshoel

Modern aircraft — both piloted fly-by-wire commercial aircraft as well as UAVs — more and more depend on highly complex safety critical software systems with many sensors and computer-controlled actuators. Despite careful design and V&V of the software, severe incidents have happened due to malfunctioning software.

In this paper, we discuss the use of Bayesian networks to monitor the health of the on-board software and sensor system, and to perform advanced on-board diagnostic reasoning. We focus on the development of reliable and robust health models for combined software and sensor systems, with application to guidance, navigation, and control (GN&C). Our …

Integrating Probabilistic Reasoning And Statistical Quality Control Techniques For Fault Diagnosis In Hybrid Domains, Brian Ricks, Craig Harrison, Ole J. Mengshoel

Ole J Mengshoel

Integrated Software And Sensor Health Management For Small Spacecraft, Johann Schumann, Ole J. Mengshoel, Timmy Mbaya

Ole J Mengshoel

Despite their size, small spacecraft have highly complex architectures with many sensors and computer-controlled actuators. At the same time, size, weight, and budget constraints often dictate that small spacecraft are designed as single-string systems, which means that there are no or few redundant systems. Thus, all components, including software, must operate as reliably. Faults, if present, must be detected as early as possible to enable (usually limited) forms of mitigation. Telemetry bandwidth for such spacecraft is usually very limited. Therefore, fault detection and diagnosis must be performed on-board. Further restrictions include low computational power and small memory.

In this paper, …

Belief Propagation By Message Passing In Junction Trees: Computing Each Message Faster Using Gpu Parallelization, Lu Zheng, Ole J. Mengshoel, Jike Chong

Ole J Mengshoel

Compiling Bayesian networks (BNs) to junction trees and performing belief propagation over them is among the most prominent approaches to computing posteriors in BNs. However, belief propagation over junction tree is known to be computationally intensive in the general case. Its complexity may increase dramatically with the connectivity and state space cardinality of Bayesian network nodes. In this paper, we address this computational challenge using GPU parallelization. We develop data structures and algorithms that extend existing junction tree techniques, and specifically develop a novel approach to computing each belief propagation message in parallel. We implement our approach on an NVIDIA …

Verification And Validation Of System Health Management Models Using Parametric Testing, Erik Reed, Johann Schumann, Ole J. Mengshoel

Ole J Mengshoel

System Health Management (SHM) systems have found their way into many safety-critical aerospace and industrial applications. A SHM system processes readings from sensors throughout the system and uses a Health Management (HM) model to detect and identify potential faults (diagnosis) and to predict possible failures in the near future (prognosis). It is essential that a SHM system, which monitors a safety-critical component, must be at least as reliable and safe as the component itself—false alarms or missed adverse events can potentially result in catastrophic failures. The SHM system including the HM model, a piece of software, must therefore undergo rigorous …

Initialization And Restart In Stochastic Local Search: Computing A Most Probable Explanation In Bayesian Networks, Ole J. Mengshoel, David C. Wilkins, Dan Roth

Ole J Mengshoel

For hard computational problems, stochastic local search has proven to be a competitive approach to finding optimal or approximately optimal problem solutions. Two key research questions for stochastic local search algorithms are: Which algorithms are effective for initialization? When should the search process be restarted? In the present work, we investigate these research questions in the context of approximate computation of most probable explanations (MPEs) in Bayesian networks (BNs). We introduce a novel approach, based on the Viterbi algorithm, to explanation initialization in BNs. While the Viterbi algorithm works on sequences and trees, our approach works on BNs with arbitrary …