Undergraduate Pilot Training Research Articles

Optimizing Undergraduate Pilot Training 2.5 at Vance Air Force Base

With a continued increase in demand for qualified pilots, in an environment of continual budget constraints, the Air Education Training Command has been searching for ways to streamline Undergraduate Pilot Training (UPT). Leadership has recently implemented UPT 2.5 to train pilots more effectively. However, the 8th and 33rd Flight Training Squadrons at Vance Air Force Base have had issues efficiently cycling students through. Although UPT 2.5 specifies that it takes no more than 90 days for students to complete the post-preflight portion of the syllabus, it currently takes 126 days on average. We developed a discrete event simulation model of pilot training to identify bottlenecks in the current system, find an optimal class size given current resources, and optimize resources to graduate 32 students per class, the number required by AETC. The results will inform recommendations to increase efficiency within UPT and reduce the pilot shortage.

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.

Proceedings of the Human Factors and Ergonomics Society 2019 Annual Meeting

We are submitting only a short abstract. We would prefer not to publish in the proceedings. This lecture discusses a quantitative methodology for demonstrating skill transfer between simulation and actual operations using a military program as a case study. Pilot Training Next (PTN) is an experimental program to increase the efficiency of Undergraduate Pilot Training (UPT) through the use of immersive, virtual-reality (VR) flight simulators. Though over a half-century of research and use has demonstrated the value of simulation, and commercial pilot training relies extensively on it, the viability of intensive VR use for military training remained an open question. PTN has provided the first steps to close this case. By substituting VR simulation for sorties, the program was able to graduate 13 out of 15 officers through the T-6 portion of training. This was accomplished in roughly one-third the time and with half as many flights as traditional UPT. Furthermore, analysts were able to model the performance data collected from this first cohort using modern, industry and academic-standard techniques. These models show that simulator performance reliably predicts skill in the actual aircraft, providing defensible evidence of skill transfer even at this intensive level of simulator reliance. They also provide credible evidence of early-warning signals that skills may not be transferring, allowing instructors to intervene and course correct if necessary. PTN-style training has the potential for great cost- and time-savings, and this first cohort provides preliminary evidence that learning quality does not have to be sacrificed to achieve this.

How Can an Ab-Initio Pilot Avert a Future Disaster: A Pedagogical Approach to Reduce The Likelihood of Future Failure

Safety is the primary concern of the civil aviation industry. This study aimed to explore the importance of learning from failures and improving the future performance of students by teaching this notion during an undergraduate course in the curriculum of the pilot training program. A questionnaire study was conducted with students of the pilot training program of the Faculty of Aviation and Aeronautical Science after the completion of a newly introduced elective undergraduate course, Aircraft Accident and Incident Investigation. Furthermore, for the first time, the contributing factors in accidents and incidents were analyzed and classified in an undergraduate pilot training program by using the Human Factors Classification and Analysis System (HFACS) as an analytical framework. The objective of the study was to find out if the “aspects of the accident investigation course at universities” were effective on “Technical Pilot Skills,” “NonTechnical Pilot Skills” and “Program-Specific Outcomes.” Correlation and Hierarchical regression analyses were conducted for this purpose. In the regression analysis, the variables were entered into the model and controlled. The results showed that the aspects of the accident investigation course improved technical and non-technical skills of the students, as well as their program-specific outcomes, and they were actively encouraged to extend their knowledge and skills beyond that required for the Airline Transport Pilot License (ATPL). It was revealed that this new undergraduate course does not only help learn the importance of non-technical skills (Crew Resource Management, Situational Awareness, etc.) but also develops and improves the technical abilities of ab-initio pilots.

Predictive Validity of Pilot Selection Instruments for Remotely Piloted Aircraft Training Outcome

Demand for remotely-piloted aircraft (RPA) support has increased dramatically over the last decade. Initial efforts to meet the demand focused on cross-training experienced manned aircraft pilots and funneling recent Specialized Undergraduate Pilot Training (SUPT) graduates to RPA pilot training. This approach reduced the number of personnel available for manned airframes and is no longer sustainable. In 2009, the USAF established an RPA career field and the Undergraduate RPA Training (URT) course to train officers with no prior flying experience to be RPA pilots. URT selection methods are very similar to those for SUPT. Some important factors for URT applicants are medical flight screening and aptitude tests [Air Force Officer Qualifying Test (AFOQT) and Pilot Candidate Selection Method (PCSM)]. The current study examined the predictive validity of the AFOQT pilot and PCSM composites for URT completion. Subjects were 139 URT students with AFOQT and PCSM scores. The training criterion was URT pass/fail and the pass rate was 74.8%. Both the AFOQT pilot (r = 0.378) and PCSM (r = 0.480) composites demonstrated good predictive validity. No minimum qualifying PCSM score exists for URT. Had a minimum PCSM score of 25 been used, the pass rate would have been 80.2%; 12 more eliminees would have been screened out compared with the current AFOQT pilot minimum qualifying score of 25. Although current selection methods are effective, based on results of several RPA job/task analyses, the Air Force is examining the utility of other measures to supplement current methods.

Pilot Candidate Selection Method

In 1993, the US Air Force implemented the Pilot Candidate Selection Method (PCSM) as an additional indicator of pilot training aptitude. PCSM scores have been shown to be related to several pilot training criteria including graduation/elimination, flying grades, class rank, and number of flight hours needed to complete training ( Carretta & Ree, 2003 ). The PCSM composite is still in use today, but the scores that contribute to it have changed. The study reported here examined the predictive validity of the current composite versus several training performance measures for 883 students attending Specialized Undergraduate Pilot Training (SUPT). Results indicated that the PCSM composite remained predictive of flying training performance despite changes in test content and flying training. After correction for multivariate range restriction and dichotomization of the graduation/elimination criterion, the correlation between the PCSM composite and T-6 completion was .53. This is consistent with the value of .46 reported by Carretta and Ree (2003) for an earlier version of the PCSM composite. Its predictiveness can be attributed to the measurement of factors that have consistently shown a relation to flying performance: cognitive ability, aviation job knowledge/experience, and psychomotor ability.

Fusing Statistical and Neural Classification for Screening Undergraduate Pilot Training Candidates

Fusing Statistical and Neural Classification for Screening Undergraduate Pilot Training Candidates

Better Candidates Fly Fewer Training Hours: Another Time Testing Pays Off

The relationship between Pilot Candidate Selection Method (PCSM) scores and the number of flying training hours required to complete United States Air Force Undergraduate Pilot Training (UPT) was investigated on a sample of 1,082 graduates. This was done to demonstrate that ability testing has real world correlates and that cost savings accrued as a result of ability testing. The criterion ‘extra flying hours’ was computed by subtracting each student's cumulative flying hours from the sample mean. The correlations (corrected for range restriction) between PCSM scores and primary and advanced flying training extra hours were −0.206 and −0.270, respectively. Demonstrating that UPT graduates with higher PCSM scores required fewer flying hours to complete training facilitates the estimation of the cost avoidance achieved by ability tests.

Basic Attributes Test: Psychometric Equating of a Computer-Based Test

The U.S. Air Force will implement a new pilot candidate selection composite based on scores from a paper-and-pencil multiple aptitude test, a computer-based test, and biographical information. A selection composite score derived from these variables will be used to indicate the probability of success in undergraduate pilot training. The purpose of this study was to equate the experimental and operational computer-based tests. The preimplementation equating was necessary (a) to determine whether the tests measured the same psychological constructs, (b) to compare the score distributions of the two computer-based tests, and (c) to develop equating tables. Equating was required so that scores from operational tests could be used in proposed pilot candidate selection equations developed on the basis of the experimental tests. Experimental and operational tests were administered to 2,034 basic recruits. Analyses were performed at the test-score and composite-score level, and several types of equatings were com...

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...

Cross Validation of Experimental USAF Pilot Training Performance Models

Abstract : A series of studies have indicated that individual differences in hand-eye coordination, information processing ability, personality and attitudes are related to USAF pilot training performance. The current investigation was designed to cross-validate these results. Eight hundred eighty-five (885) USAF Undergraduate Pilot Training (UPT) students were divided randomly into two groups. Pilot selection models that used a combination of Air Force Officer Qualifying Test (AFOQT) and Basic Attributes Test (BAT) battery scores were developed independently for each group and then cross-validated with the other group. In the model development phase, subjects with good hand-eye coordination who made quick decisions were more likely to complete UPT successfully in both groups. Although there was some reduction in the validity coefficients in the cross-validation phase, the selection models were related significantly to UPT final outcome in both groups. These results suggest that the AFOQT/BAT pilot selection models are sufficiently robust to be used as adjuncts to operational USAF pilot trainee selection procedures. Keywords: Aptitude testing, Basic attributes test, Pilot selection.

Air Force Pilot Personality: Hard Data on the "Right Stuff".

Three-hundred and fifty Air Force pilots undergoing Undergraduate Pilot Training were administered the Personality Research Form (PRF) and the Millon Clinical Multiaxial Inventory (MCMI) within the first four weeks of training. Pilots were significantly different from non-flying college students on several PRF scales. Cluster analyses indicated that three very distinct personality types exist in the data. These types were supported through multiple cross-validations. Descriptions of pilot personality types are given in terms of PRF and MCMI personality variables. The accuracy of prevailing stereotypes is examined in light of the data. Implications for future research and pilot selection criteria are discussed.

A Simulator-based Approach to Training In Aeronautical Decision Making

The effectiveness of a simulator-based approach to training pilot skills in risk assessment and decision making was evaluated in a sample of pilots enrolled in a university aviation science program. The 16 experimental group subj ects received four hours of classroom instruction designed to enhance pilot judgment skills. followed by four simulated cross-country flights during which several critical inflight events occurred. Subjects in the control group received classroom instruction in basic instrument flying. followed by simulator sessions emphasizing instrument flight. Measures of pilot judgment were obtained on all subjects before and after the training and subjects in the experimental judgment-trained group performed significantly better on the post-training simulation than did control group subjects. The findings suggest that significant gains in pilot decision-making skill can be obtained through the use of the judgment training materials along with simulator practice. The implications of Air Force undergraduate pilot training are discussed.

An Integrated Approach to Pilot Performance Evaluation

The issue of norm- versus criterion-referenced evaluation methods is discussed in the context of undergraduate pilot training, and an approach that integrates both methods is described. It is suggested that such an integrated approach would facilitate advanced training management concepts such as proficiency advancement, but would not pose significant user acceptance difficulties.

Time-Delay Problems Encountered in Integrating the Advanced Simulator for Undergraduate Pilot Training

An advanced Computer Image Generation (CIG) visual system has been integrated with an advanced Flight Training Research Simulator. The integration design was the first developed for integrating a CIG visual system with a sophisticated flight simulator. There was much concern for the unique CIG system transport delay, and techniques were developed which proved to be quite successful in compensating for the majority of this delay. However, not enough concern was given to previously unrecognized and unreported excessive motion system delays which were encountered during final integrated system tests. The integration scheme and the impact of iteration rates, visual and motion system delays, and delay compensation on visual and motion cue coordination as perceived by pilots are presented.

Behavioral Taxonomy of Undergraduate Pilot Training Tasks and Skills

The analysis and specification of fundamental flying abilities which comprise the training objectives of Air Force undergraduate pilot training (UPT) was performed. The taxonomy of UPT tasks and skills is an analytical tool of considerable generality that can be used to aid in understanding the essential requirements of flying training. Surface analyses of fifty UPT maneuvers generated task element descriptions subdivided into a series of cue, mental action, and motor action sequences. The resulting task information was used to identify the pilot skills required to execute the flying tasks described. A set of classification rules organized the skills into a taxonornic cubic concept in which cues, mental actions, and motor actions represented cube faces. The classification rules were validated empirically and used to verify all surface analyses. The required task skills were then organized into a matrix system for simple data retrieval operations.

Advanced Instructional Provisions and Automated Performance Measurement

An advanced simulation research system is being developed to support experiments oriented toward quantifying the effect on transfer of training of alternative training and simulation techniques. Included in the system is sophisticated training research software which automates most of the functions traditionally performed by flight simulator instructors or operators. Also under development are techniques for automatically assessing pilot proficiency in the simulator and in the operational aircraft. Recent feasibility studies in performance measurement resulted in identification of necessary flight variables for assessing proficiency on two undergraduate pilot training maneuvers; established alternative approaches to developing measurement techniques on a broader scale; applied existing technology to develop an in-flight data acquisition system for the T-37 aircraft; and raised (or, more correctly, reraised) issues regarding standardization in instruction and rating procedures.