Dynamic operator interaction with multi-modal automation systems

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Aeronautics and Astronautics, 2011.This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.Cataloged from student-submitted PDF version of thesis.Includes bibliographical references (p. 103-114).The ability of operators to "gracefully transition" (maintaining control and awareness of the system without excessive workload or decrements in flight performance) between levels of automation (LOA) in several case studies and in a simulated lunar landing was investigated in anticipation of future lunar missions. Endsley's situation awareness model (extended to apply to supervisory control systems) and the Sheridan/Verplank and Proud/Hart LOA scales were used to analyze six maritime, aviation, and aerospace case studies and formulate design guidelines for enhancing mode transitions. These motivated an experiment in which thirteen subjects with flight simulator experience flew 24 approach trajectories (half including a landing point redesignation) that transitioned from a fully automatic flight control mode to either: pitch rate-control/attitude hold (RC/AH) with automatic rate-of-descent (ROD), roll-pitch-yaw (RPY) RC/AH with automatic ROD, or RPY RC/AH with incremental ROD. Subjective and objective workloads were measured using a Modified Bedford Scale and secondary task response time, respectively. A tertiary task - verbal callouts of altitude, fuel, and location, provided a measure of pilot situation awareness. Flight performance was evaluated using the pitch axis tracking error. Friedman pairwise tests demonstrated that secondary task response time significantly increased following the mode transition. Subjects' workload ratings, when ranked, showed unanimous agreement that workload was lowest prior to the transition, and highest during. The accuracy of the situation awareness verbal callouts decreased significantly after the transition. The immediate effect of redesignation was statistically concordant across subjects. Pitch axis tracking mean square error following a mode transition was greater in trials with redesignations (p = 0.0005), and increased consistently with control mode difficulty (p = 0.025) in runs with no redesignation, but not in runs with redesignations. Using callouts to assess the dynamics of situation awareness is a novel technique. Dramatic changes in subjective and objective workload and situation awareness occur after mode transitions, depending on control mode difficulty, that have an apparently reciprocal relationship. The case studies and experimental results suggested a dozen guidelines for design of supervisory control systems intended to promote transition gracefulness.by Christopher James Hainley, Jr.S.M