Assessing Human Task Performance When Performing Electronic Procedures

Period of Performance: 01/01/2011 - 12/31/2011


Phase 1 SBIR

Recipient Firm

Traclabs, Inc.
100 North East Loop 410, Suite 520
San Antonio, TX 78216
Firm POC
Principal Investigator


Limited astronaut time can result in high crew workload and fatigue during International Space Station (ISS) operations. Additionally crew skills can decline over extended mission duration onboard ISS. These stressors can impact task performance both in terms of increased errors and decreased efficiency. As humans go deeper into space, these issues will become even more pronounced. Methods are needed to monitor tasks as they are performed and to detect degraded task performance. TRACLabs proposes to develop algorithms and software for monitoring human performance of procedures. Our approach will integrate performance measuring software with procedure software and the displays used to execute procedures. We will build on prior work by TRACLabs on the Procedure Representation Language (PRL) and our agent software that aids humans interacting with PRL procedures. The performance measuring software will use techniques developed at the University of Pittsburgh that monitor keystrokes and mouse utilization to compute the speed and accuracy of individual pointing and text entry actions and adherence to procedure definitions. We will adapt techniques for measuring task performance of the disabled to measuring task performance of the able-bodied when stressed. We propose that performance changes due to these situational disabilities can be detected in the same way as performance changes due to physical disabilities. These techniques are particularly attractive for use at NASA because they are non-invasive and do not require additional equipment to implement. Performance measures will be computed in real-time as tasks are performed and thus will be available for use during task performance. Procedure information will guide the collection and interpretation of the human task performance data.