Abstract: | The lack of aircraft state awareness has been one of the leading causal and contributing factors in aviation accidents. Many of these accidents were due to the flight crew’s failure to understand the automation modes and properly monitor the aircraft energy and attitude state. The capability of providing flight crew with improved airplane state awareness (ASA) is essential in ensuring aviation safety. Predictive alerting methods achieve improved ASA by integrating onboard information to estimate and subsequently predict the aircraft state based on: (i) aircraft state related information output by the onboard avionics, (ii) the aircraft configuration, (iii) appropriate aircraft dynamics models of both the active modes and the modes to which can be transitioned via simple pilot actions, (iv) and accurate models of the uncertainty of the dynamics and sensors. Onboard avionics inputs include measurements from onboard navigation systems such as global navigation satellites systems (GNSS), inertial navigation systems, and air data. This paper focuses on evaluating the sensitivity to sensor uncertainty of energy state prediction performance and the ability of the system to provide reliable alerts based on these predictions. Specific predictive alerts include the prediction of: (a) stall and overspeed conditions, (b) high-and-fast conditions, (c) unstable approach conditions, (d) and automation mode transitions. This paper provides a detailed description of the prediction algorithms and predictive alerting display concepts. It furthermore shows results of the proposed methods. This analysis makes use of flight data collected during a recent NASA flight simulator study in which eleven commercial airline crews (22 pilots) completing more than 230 flights. Intentional uncertainty was introduced to the sensor inputs and the outputs were evaluated in terms of: missed detections of future hazardous situations, missed timely predictive alerts, and false alerts. |
Published in: |
Proceedings of the ION 2017 Pacific PNT Meeting May 1 - 4, 2017 Marriott Waikiki Beach Resort & Spa Honolulu, Hawaii |
Pages: | 87 - 99 |
Cite this article: | de Haag, Maarten Uijt, Engelmann, James, Mourning, Chad, Duan, Pengfei, "Keynote: Evaluation of Energy State Prediction and Predictive Alerting Methods under Sensor Uncertainty," Proceedings of the ION 2017 Pacific PNT Meeting, Honolulu, Hawaii, May 2017, pp. 87-99. https://doi.org/10.33012/2017.15040 |
Full Paper: |
ION Members/Non-Members: 1 Download Credit
Sign In |