Use this URL to cite or link to this record in EThOS:
Title: Dynamic reconfiguration of safety-critical systems : automation and human involvement
Author: Montano, Giuseppe
ISNI:       0000 0004 2715 8692
Awarding Body: University of York
Current Institution: University of York
Date of Award: 2011
Availability of Full Text:
Access from EThOS:
Access from Institution:
This thesis describes the design and evaluation of a novel Decision Support System (DSS) for naturalistic, safety-critical decisions on-board modern aircraft. The system is intended to improve pilots’ decision-making accuracy and performance, by supporting human cognitive strategies. In recent years, the development of dynamically reconfigurable Safety-Critical Manned Systems (SCMS) has acquired increasing attention in several engineering domains including civil and military aerospace, marine and ground transportation. Dynamic reconfiguration of the on-board control systems enables adaptation to the changing conditions during operation. At the occurrence of a fault or damage, reconfiguration allows for the transition to a degraded operating mode by deactivating a number of services in order to preserve sufficient resources for the provision of essential functionality. The current focus of mainstream research is on full autonomy and full authority solutions, which nonetheless make the transition to a degraded mode transparent to the operator, as much as possible. This thesis takes a different approach, developing a human-centred perspective: by drawing on well-established fields such as Cognitive Engineering, Human Factors and Naturalistic Decision Making, it identifies limitations of fully automated dynamic reconfiguration solutions, including some safety problems, and proposes novel technology to keep the operator much more effectively “in the control loop” during reconfiguration. A review of the relevant literature leads to the identification of three main research problems: (a) determining the characteristics of effective decision support information for SCMS dynamic reconfiguration decisions; (b) developing DSS technology to autonomously generate the type of information required; (c) developing a methodology to evaluate and validate the performance of the DSS and assess its effectiveness in support of the decision making activity. First, pilot behaviour during fault management decisions is investigated and a novel design for decision support information that parallels human cognitive strategies is devised. The hypothesis advanced is that decision support information that favours mental simulation by including (a) explanations that justify each reconfiguration alternative, (b) implications for each alternative and (c) an assessment of the uncertainty embedded in the sensor information would have a positive impact on both human decision accuracy and performance. Second, a novel Constraint-based DSS is developed to generate the type of information suggested by the research hypothesis. A number of algorithms and software applications designed to handle the reconfiguration process and generate decision support information are developed and their performance is assessed. The tools developed are integrated into the Safe and Interactive Reconfiguration Architecture (SaIRA), a novel framework for automated decision support. Third, seven experiments, which involved thirteen civilian aircraft pilots, were performed to (a) empirically verify the claims advanced throughout the thesis concerning the issues with automation and human involvement during SCMS dynamic reconfiguration, and (b) to assess the effectiveness of SaIRA. A validation methodology that merges a number of relevant objective and subjective metrics is proposed. The experiments reveal that SaIRA improves pilots’ decision accuracy, decision performance, situation awareness and, more generally, their cognitive readiness whilst reducing cognitive workload and frustration under heavy time pressure. Whilst this work has been undertaken in the context of civil aviation systems, there is reason to believe such classes of decision support system would be of much wider applicability.
Supervisor: McDermid, John Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available