Air Force Pilot Training Research Articles

Predicting success in United States Air Force pilot training using machine learning techniques

The chronic pilot shortage that has plagued the United States Air Force over the past three years poses a national-level problem that senior military members are working to overcome. Unfortunately, not all pilot candidates successfully complete the necessary training requirements to become fully qualified Air Force pilots, which wastes critical time and resources and only further exacerbates the pilot shortage problem. Therefore, it is important for the Air Force to carefully consider whom they select to attend pilot training. This research examines historical specialized undergraduate pilot training (SUPT) candidate data leveraging a variety of machine learning techniques to obtain insights on candidate success. Computational experimentation is performed to determine how selected machine learning techniques and their respective hyperparameters affect solution quality. Results reveal that the extremely randomized tree machine learning technique can achieve nearly 94% accuracy in predicting candidate success. Additional analysis indicates degree type and commissioning source are the most important features in determining candidate success. Ultimately, this research can inform the modification of future SUPT candidate selection criteria and other related Air Force personnel policies.

Vulnerabilities of the Romanian Air Force pilot training system manifested in the flight training process

The vulnerability of a flight instruction system is understood as a state of affairs, processes or phenomena in the organization’s internal life that diminishes its ability to react to existing or potential changes. Starting from the results of previous research conducted by the authors, the article aims to highlight the need for new ways to train Air Force pilots and suggests possible solutions. The purpose of this study is thus to identify the areas of vulnerability proper to the organization providing the training by synthesizing survey results based on the subjects’ professional experience. Accordingly, we deemed necessary and at the same time pertinent to consider the opinions of the staff involved in the practical training process. For this reason, the evaluation carried out between June and July 2018 involved the Aviation Training Group at “Aurel Vlaicu” Air Force Base.

Alternobaric vertigo: prevalence in Portuguese Air Force pilots

Conclusion. Having found a prevalence rate of alternobaric vertigo in Portuguese Air Force pilots that is somewhat higher than previously reported, we underline the importance of implementing education on the management of this condition as part of routine Air Force pilot training programs. Objectives. Alternobaric vertigo is a condition in which transient vertigo with spatial disorientation occurs suddenly during flying or diving activities, caused by bilateral asymmetrical changes in middle ear pressure. Its prevalence is very likely underestimated and under-reported, with the 10–17% prevalence rate mentioned in early literature not being challenged by recent data. Subjects and methods. To assess its actual prevalence, the authors requested all high performance aircraft pilots presently on active duty in the Portuguese Air Force to anonymously answer a questionnaire on alternobaric vertigo symptoms, after a short briefing on the subject. Results. A 29% prevalence rate of in-flight episodes consistent with alternobaric vertigo was obtained.

Classification models for sequential flight test results for selecting air force pilot trainee

The main purpose of this paper is to present the selection criteria for Air Force pilot training candidates in order to save costs involved in the three stage sequential pilot training procedures currently used in Korea. We use classification models such as decision tree, logistic regression and neural network based on the aptitude test results of 288 ROK Air Force applicants in 1994-1996. Various classification models are compared in terms of classification accuracy, Receiver Operating Characteristic chart and Lift chart. As a result, neural network is evaluated as the best classification model for each sequential flight performance while logistic regression model outperforms the others for the overall flight result. The fitted logistic regression indicates that the factors such as attention sharing, instrument reading, and mechanical comprehension having significant effects on the flight results. We expect that the use of such classification models can increase the effectiveness of flight resources.

단계별 비행훈련 성패 예측 모형의 성능 비교 연구

The main purpose of this paper is to present selection criteria for ROK Airforce pilot training candidates in order to save costs involved in sequential pilot training. We use classification models such as Decision Tree, Logistic Regression and Neural Network based on aptitude test results of 288 ROK Air Force applicants in 1994-1996. Different models are compared in terms of classification accuracy, ROC and Lift-value. Neural network is evaluated as the best model for each sequential flight test result while Logistic regression model outperforms the rest of them for discriminating the last flight test result. Therefore we suggest a pilot selection criterion based on this logistic regression. Overall, we find that the factors such as Attention Sharing, Speed Tracking, Machine Comprehension and Instrument Reading Ability having significant effects on the flight results. We expect that the use of our criteria can increase the effectiveness of flight resources.

Pilot safety enhancement for on‐board ground collision avoidance systems

The goal of the paper is to present an enhancement of the existing on‐board ground collision avoidance system (GCAS) that is designed to increase pilot safety in USAF A‐10 aircraft. The A‐10 is a single‐seat, twin‐engine aircraft with a 30mm, seven‐barreled Gatling gun and 11 weapon pylons designed to fly at low level in close air support missions. The GCAS system provides both visual and aural cues for a pilot‐initiated recovery. The proposed algorithm of GCAS enhancement is built on a simple linear regression model that predicts the recovery height of the aircraft following a warning call and allows pilots to compare their own training events with flight test standards. This paper presents a discussion of model development, validation and comparison of the model predictions with actual flight test events. A comparison of recovery techniques and pilot options is included. A series of recommendations and possible usage for Air Force pilot training are also discussed.

Pilot-Candidate Selection Method: Sources of Validity

Six hundred seventy-eight Air Force pilot training candidates were tested with a paper-and-pencil aptitude battery and computer-administered tests of psychomotor skills, information processing, and attitude toward risk. A self-report of flying experience was also collected. These data were used in regression analyses to determine which variables provided the best prediction of two flying criteria: pass-fail flying training, and class rank at the end of flying training. The paper-and-pencil tests were found to be the best predictors. The measures of flying experience, psychomotor skills, and attitude toward risk incremented the prediction of the criteria above the prediction provided by the paper-and-pencil tests by 23%. Computer-administered information-processing measures were not found to be incremental to the other variables in the prediction of the criteria.

Understanding the Relations Between Selection Factors and Pilot Training Performance: Does the Criterion Make a Difference?

Although several valid predictors of U.S. Air Force pilot training performance have been identified, it was felt that our understanding of the relation between preselection personnel attribute data and training performance was limited by the dichotomous nature of the outcome indicators (i.e., graduation or elimination) and by the disproportionate number of people in the outcome categories. A weighted composite of Undergraduate Pilot Training (UPT) flying performance data (i.e., daily flying grades, check flight grades, and academic grades) was shown to be related closely to advanced training recommendations (fighter vs. nonfighter aircraft). This suggested that this composite score was a reasonable measure of pilot candidate quality because fighter aircraft assignments are considered more prestigious and demanding than nonfighter assignments. When the weighted composite algorithm was modified to include UPT eliminees, however, it demonstrated little utility in adding to our understanding of the relation b...