Automated fault-management in a simulated spaceflight micro-world

Area of Science:

• Human-Computer Interaction
• Spaceflight Operations
• Automation and Control Systems

Background:

• Extended human spaceflight necessitates increased onboard autonomy and automated fault management (FM).
• Operator reliance on automated systems raises human factors concerns: cognitive skill, situational awareness (SA), trust, and workload.
• This study investigates human performance with intelligent FM support for an autonomous atmospheric control system.

Purpose of the Study:

• To examine the human performance consequences of operator interaction with intelligent fault management support.
• To evaluate the impact of different levels of automation (LOA) on operator performance during system faults.
• To assess the role of automation in maintaining situational awareness and cognitive workload.

Main Methods:

• An expert system provided fault management support at low (guide), medium (advisory), and high (implementation) levels of automation (LOA).
• Operators interacted with an autonomous, space-related, atmospheric control system.
• Ten percent of trials included complete failure of fault management support to assess system robustness.

Main Results:

• Automation improved diagnostic accuracy, reduced fault identification time, and lowered subjective operator workload.
• Fault identification times worsened more at medium than high LOA during automation failure.
• High LOA operators, freed from recovery tasks, engaged more in fault assessment activities during reliable automation.

Conclusions:

• High-level automation does not necessarily threaten situational awareness (SA); it can enhance fault comprehension.
• Automatic implementation of recovery plans at higher LOA frees cognitive resources for better fault understanding.
• Effective automation interface design is crucial for supporting efficient information sampling and maintaining SA.