Machine-controlled adaptive training: task difficulty varies as a function of trainee performance¶

Theory · All theories
Evidence · 3 claims (2 for, 1 against) · 1 study (1 causal), q2 · 0 of 1 report an effect size · 3 claims rest on one study

Description¶

In adaptive training, the task presented to the trainee varies as a function of how well he performs. In machine-controlled training, adaptive logic performs a function analogous to that performed by a skilled instructor, adjusting task difficulty based on measured performance. The concept was demonstrated for flight training and was being implemented in the 2B24 synthetic flight training system for helicopters. This study tested the concept on a single-axis compensatory manual tracking task.

Design Implications¶

Context¶

Requirements¶

• A performance measure must be available to drive the adaptive logic; the article notes speed of response is affected by the method of performance measurement

Constraints¶

• Several problems remain to be solved before the concept can be utilized to its fullest potential, including the choice of adaptive variables

Target Learners¶

• trainee pilots
• novice manual-control trainees

Target Learning Objectives¶

• manual control skill
• compensatory tracking skill

Claims¶

• Gain-effective time constant product is slightly superior to forcing function amplitude as an adaptive variable [+W]
• Manual adaptation is slightly superior to automatic adaptation in adaptive training of manual control [+W]
• System compensation as implemented is not a satisfactory adaptive variable [-W]

Related Theories¶

• Effective time constant as a quantifiable construct predictive of task difficulty

Examples¶

• Use shaped-spectrum filtered error for rapid adaptive logic response

Key Sources¶

• Norman, D. A.; And Others. (1972). Adaptive Training of Manual Control: 1. Comparison of Three Adaptive Variables and Two Logic Schemes. Life Sciences, Inc. https://eric.ed.gov/?id=ED059585 doi:10.1037/e426822004-001