Back to Conference Program

Thursday 4 November 2010 — 09:30-11:00

Moderator: Michael Feary, Ph.D., NASA, U.S.A.

Panelists:

Randall Mumaw, Ph.D., Boeing, U.S.A.

Philippe Palanque, Ph.D., University of Toulouse, France

Asaf Degani, Ph.D., General Motors, Israel

Denis Javaux, Next Step Solutions, Belgium

ABSTRACT

As aerospace operations in the air and on the ground become more dependent on automation, new methods are needed to

enable improved evaluation of human automation interaction in the design and evaluation process. New airworthiness

regulations and regulatory certification processes are beginning to focus on the design, verification and testing of the pilotautomation interaction. However, current human computer interaction analysis methods and tools need improvement to be

usable and useful within the constrained timeline of real world design and certification processes. It is acknowledged that

improvements are needed both with the tools and the underlying fundamental and theoretical work underlying the

tools (e.g. defining relevant metrics), to provide designers and regulators with the means of testing and providing useful

feedback about the efficiency and effectiveness of these interactions. This panel will discuss the potential areas of focus,

and possible first steps for improving the evaluation methods and tools needed to identify Human – Automation Interaction risks early in the design process.

OVERVIEW

With the growing increase in the complexity of modern systems and increased flexibility afforded by modern, computer-driven interfaces, the need for sound design methodologies and evaluation/analytical tool is on the rise. The application of

automation in the design of modern systems found in the aviation, automotive, medical, and other high risk domains

necessitates a more detailed understanding and analysis of the relation between the information provided on the interface and

the actual working of the machine. Traditional human factors methods such as controlled experiments with human subjects and a variety of evaluation methodologies are not suited to address the problem of how to make sure that the design of user

interfaces and user interaction is free of faults and is correct. Rigorous and systematic methods of analysis are needed to

consider possible configurations of system states and make sure that user interaction with these configurations is consistent with the design requirement and that faulty states are avoided. The panelists will discuss the state of the art of these approaches and the challenges that lie ahead to providing useful tools. Particular discussion points may include:

1. •The ability to provide methods and tools that address specific Human – Automation interaction performance issues during the design process.
   

2. •The ability to provide the design community with tools that are usable early in the design process.
   

3. •The ability to provide the design community with tools that can compare different automation and interface concepts quickly for a given task.
   

4. •The ability to provide the design community with methods and tools that can predict operators errors before they actually occur and thus support the design of error-tolerant human-automation interactions.
   

5. •The state of the art in measuring relevant performance metrics, going beyond Time – on Task.
   

6. •The state of the art in methods that appropriately address complexity and large state spaces.
   

The panel will also discuss the next steps to helping the design community address the new regulatory requirements (e.g. CS/AMC 25.1302[1]) and the next wave changes introduced by next generation aviation operations.