Human performance in a multiple-task environment: effects of automation reliability on visual attention allocation.

Area of Science:

• Human-Computer Interaction
• Cognitive Psychology
• Automation Science

Background:

• Diagnostic automation is intended to reduce workload in complex environments.
• Previous research shows mixed results regarding the effectiveness of automation in aiding performance.
• Understanding attention allocation is crucial for evaluating automation's impact.

Purpose of the Study:

• To investigate how automation influences visual attention allocation in a multiple-task environment.
• To examine the effects of varying automation levels and reliability on user attention.
• To assess the consequences of automation removal on task performance and attention.

Main Methods:

• Developed a simulated multiple-task environment to test automation effects.
• Manipulated automation levels and reliability during user interaction.
• Measured attention allocation by tracking task view frequency.
• Assessed performance after transferring users to a no-support system.

Main Results:

• Participants with automation allocated attention based on task frequency.
• Tasks that received the most benefit from automation were most negatively impacted upon its removal.
• Automation failure showed an inverse relationship between support provided and performance cost.

Conclusions:

• The impact of automation on multiple-task performance is contingent on the specific attributes of the tasks involved.
• Automation can lead to detrimental effects on performance for highly supported tasks when support is withdrawn.
• Careful consideration of task characteristics is necessary when designing and implementing automation systems.