An Intelligent Interface Construction and Evaluation Assistant

We propose the development of an innovative software tool that will permit payload developers with no human factors training to produce interfaces that are well suited to the tasks required by the payload, and that conform to NASA commonality standards. Developing such a tool represents a significant challenge; thus, we suggest an integrated approach drawing upon a range of AI techniques. An intelligent, visually-based knowledge elicitor will guide the developer through the task capture and analysis process. The resulting task model and a battery of intelligent design advisors, assisting in everything from widget selection and initialization to graphical element distribution, will allow the system to collaborate with the developer in constructing a task-based interface. The resulting interface will be analyzed as a whole by design critics drawn from human factors research and NASA commonality standards. The developer will then be able to run usability testing from within the same framework, and the tool will help him analyze the testing results. Finally, the developer will be able to export the interface in code form for continued application development. We proved the feasibility of our approach with a Phase I prototype and propose the implementation of a full-scale payload interface development system.