As an objective evaluator of psychomotor skills, the automated pilot selection system (APSS) gives a reliable indication of whether a candidate possesses the aptitudes and dexterity needed to become a pilot. The APSS significantly reduces the costs associated with pilot training failures by allowing only those candidates who attain a passing score on the APSS exercises to proceed to primary flight training. At the Canadian Aircrew Selection Centre, the cost saving is estimated to be more than 15 percent of the cost of training without the APSS.

The APSS is the result of over ten years' close collaboration between Atlantis and the Canadian Forces Aircrew Selection Centre (CFASC) in developing and validating the prediction algorithms. The APSS predictions are based on norms established from data collected on over 400 pilot training candidates.

The APSS supplied by Atlantis to the Air Command of the Canadian Forces includes five APSS cockpit units and a Data Analysis Centre (DAC). The cockpit units are mounted on a 360-degree motion base and include a digital audio-visual system. All five APSS cockpit units can be operated by a single operator seated at the DAC.

The DAC performs score calculations and maintains the data norms on which the pass/fail predictions are based. The DAC allows the instructor to:

• Monitor the operational status of all APSS units
• Monitor the candidates' progress
• Create and modify the syllabus
• Analyse candidate performance data
• Print candidate performance data.

HOW THE APSS WORKS

The candidate, who is assumed to have no experience with aircraft or the theory of flight, is placed in the APSS cockpit, which closely resembles the actual training environment. The candidate then performs a series of exercises that cover the operation of a single-engine aircraft. These exercises contain flight tasks and manoeuvres structured to test the candidate's basic psychomotor abilities, learning rate, multiple task integration skills, and performance under task overload conditions. The complexity of the exercise increases as the candidate progresses.

The APSS collects data on various aspects of the candidate's performance every half-second and performs predictive calculations on this data at the end of each session. The prediction calculations are based on norms compiled from data on over 400 pilot training candidates. The actual pass or fail thresholds used for each session can be adjusted.

One of the key features of the APSS is its ability to add the results of the norm calculation for each candidate to the norm data file. This results in a self-correcting system that has a norm for all candidate types.