Related Experiment Videos
Level of automation effects on performance, situation awareness and workload in a dynamic control task
1Department of Industrial Engineering, Mississippi State University 39762-9542, USA.
Ergonomics
|February 27, 1999
Summary
Exploring levels of automation (LOA) in dynamic tasks shows human operators benefit most when automation handles implementation, but this is risky if the system fails. Joint human-machine option generation harms performance.
Area of Science:
- Human-Computer Interaction
- Automation and Control Systems
- Cognitive Psychology
Background:
- Traditional automation models often use binary function allocation (human or machine).
- Intermediary levels of automation (LOA) aim to maintain operator engagement, improving situation awareness and reducing out-of-the-loop issues.
- A new LOA taxonomy is proposed for diverse psychomotor and cognitive tasks.
Purpose of the Study:
- To present a novel LOA taxonomy.
- To assess the impact of this taxonomy on human/system performance, situation awareness, and workload in a dynamic cognitive control task.
- To investigate the effects of automation failures on operator performance.
Main Methods:
- Developed a LOA taxonomy categorizing human/computer control functions (monitoring, option generation, selection, implementation).
- Assessed the taxonomy's impact through simulation trials with 30 subjects performing a dynamic cognitive control task.
- Measured human/system performance, situation awareness, workload, and out-of-the-loop performance decrements, including during automation failures.
Main Results:
- Automation of task implementation yielded the greatest performance benefits under normal conditions.
- Operator removal from implementation negatively impacted performance recovery after automation failures.
- Joint human/system option generation degraded performance compared to single-entity generation.
- Automating decision-making (option selection) reduced workload and increased situation awareness, with only minor performance gains.
Conclusions:
- LOA design is critical; automating implementation offers benefits but poses risks during failures.
- Human-system interaction in option generation needs careful consideration to avoid performance degradation.
- Automating decision-making can enhance operator state (awareness, workload) but requires careful balancing with performance outcomes.