Cognitive Aspects of Automated Target Recognition Interface Design: An Experimental Analysis
Abstract
The focus of this research was on the intersection of cognitive and perceptual aspects of human target recognition performance, and on potential enhancements of the human-ATR interface. Three series of experiments were conducted with active duty Army pilots. Each study attempted to lay a scientific basis, and to test a practical methodology, for a promising ATR design application. The studies address the following issues in ATR-human interface design: (1) effective displays of target classification conclusions to support rapid verification and application to the mission (2) effective displays of target imagery to support rapid and accurate user verification of ATR conclusions, and (3) effective support for decision making processes that allocate user attention, decide where and how long to verify ATR conclusions, and determine which targets to engage. Our results suggest that: (1) ATR conclusions should be labeled at different levels of specificity for different types of vehicles; (2) enhancement of vehicle profile and selected vehicle details can improve speed and accuracy of visual recognition; and (3) engagement decision making is improved by techniques for quickly guiding user attention to images classified as high-confidence enemies, high confidence friends, or significant and uncertain.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 14, 1997
- Accession Number
- ADA530971
Entities
People
- Bryan Thompson
- Jared T. Freeman
- Marvin S. Cohen