Identifying Requirements for Effective Human-Automation Teamwork
Previous studies have shown that poorly designed human-automation collaboration, such as poorly designed communication protocols, often leads to problems for the human operators, such as: lack of vigilance, complacency, and loss of skills. These problems often lead to suboptimal system performance. To address this situation, a considerable amount of research has been conducted to improve human-automation collaboration and to make automation function better as a “team player.” Much of this research is based on an understanding of what it means to be a good team player from the perspective of a human team. However, the research is often based on a simplified view of human teams and teamwork. In this study, we sought to better understand the capabilities and limitations of automation from the standpoint of human teams. We first examined human teams to identify the principles for effective teamwork. We next reviewed the research on integrating automation agents and human agents into mixed agent teams to identify the limitations of automation agents to conform to teamwork principles. This research resulted in insights that can lead to more effective human-automation collaboration by enabling a more realistic set of requirements to be developed based on the strengths and limitations of all agents.
- Research Organization:
- Idaho National Lab. (INL), Idaho Falls, ID (United States)
- Sponsoring Organization:
- DOE - NE
- DOE Contract Number:
- DE-AC07-05ID14517
- OSTI ID:
- 1149012
- Report Number(s):
- INL/CON-14-31340
- Resource Relation:
- Conference: Probabilistic Safety Assessment & Management conference,Honolulu, HI, USA (Sheraton Waikiki),06/22/2014,06/27/2014
- Country of Publication:
- United States
- Language:
- English
Similar Records
A New Appraisal- Lessons from the History of Efforts to Value Green and High-Performance Home Attributes in the United States
A non-cooperative meta-modeling game for automated third-party calibrating, validating and falsifying constitutive laws with parallelized adversarial attacks