Automation Effects in the Cockpit: A Low-Fidelity Investigation

Abstract

The effects of automation and task difficulty on flight performance, subjective and objective workload, and a problem solving task were investigated in a low fidelity flight simulator. Forty-eight, two-person crews flew two forty-five minute scenarios that required the crew to select and obtain relief supplies for delivery to a disaster site. Two levels of automation (i.e., presence or absence of an autopilot) and two levels of task difficulty (i.e., presence or absence of wind and turbulence) were combined to yield a 2 × 2 design. Twenty-four crews performed in both levels of automation and one level of task difficulty. Results indicated that although crews in the automated condition reported less subjective workload, only one of the three measures of flight performance was affected by automation. In contrast, objective workload, as measured by performance of a secondary task, was increased for the pilot in the automated condition. In addition, under high task difficulty, problem solving was worse m the automated condition than in the manual condition. The results are discussed in terms of their support of earlier hypothesized effects of automation in the cockpit.