Open Collections

Evidence-based information fusion platform for estimation of forest fire pilot fitness for duty

University of British Columbia

Over the last fifteen years there have been numerous examples of extreme forest fires in British Columbia and Alberta alone. These massive, quick moving fires require strategic air and ground maneuvers in an effort to contain and/or redirect the fires. Currently, the hours of work and rest periods for forest fire pilots are determined by Transport Canada and fall under the same regulations as commercial airline pilots called the Canadian Aviation Regulations (CARs). However, aerial operations, such as forest fire pilots, have vastly different work conditions, scheduling and workload throughout the flight. Therefore, the focus of this study was on the development of an evidence-based information fusion platform which estimates the pilots’ fit for duty index to enhance the decision makers’ and planners’ ability to assess their pilots. Fourteen Conair pilots participated in this study over the 2018 wildfire season. Subjective, cognitive, physiological and sleep measurements were collected each day for the analysis of the pilots’ fatigue and workload throughout the 2018 fire season. The evidence-based information fusion platform was developed using fuzzy logic and Dempster-Shafer theory of evidence in MATLAB and RStudio. This platform fuses the collected data throughout a pilot’s shift to determine the pre and post shift fit for duty index values. Fuzzy logic and Dempster-Shafer theory of evidence were implemented to manage vague and subjective measurements related to the human physical and mental state and handle any conflicting evidence and measurement uncertainty, respectively. The developed evidence-based information fusion platform successfully provided individualized color-coded fit for duty indices based on the pilots’ collected data through a robust and adaptive user interface. This platform can be used as a guide to monitor and manage the pilots’ fitness for duty for decision making and scheduling. The platform was developed based on a small sample of pilots over one wildfire season. Further data collection, analysis and verification throughout additional fire seasons is required to improve the robustness of the fit for duty index estimation and adopt it in practice.

Text

2019-04-01

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/69386

Mechanical Engineering

University of British Columbia

UBCO

http://creativecommons.org/licenses/by-nc-nd/4.0/