 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
Human Performance Engineering Practice
Case Studies:
Extending and Evaluating Human Performance Simulation
with COGNET/iGEN
________________________________________________Challenge
Systems engineering, training, and work design could be greatly improved by computer models that realistically simulate how humans work. These models could apply skills and reasoning but also commit errors and encounter time or accuracy constraints, just as humans would. However, complete, validated models don't yet exist.
The available human performance models needed to be benchmarked against real human behavior and extended to model multi-tasking and learning.
The Air Force Research Laboratory’s long-range research project, Agent-based Modeling and Behavioral Research, performed the benchmarking and extension for human performance models, including the CHI Systems COGNET/iGEN product.Result
CHI Systems Human Performance Engineering Practice and Cognitive Engineering Practice, working with additional support of the Office of Naval Research, collaborated to extend iGEN’s modeling capabilities to:
- make better predictive simulations for the time aspects of both physical and cognitive activities
- provide a representation of metacognitive self-awareness so that the model can better monitor and adapt is own behavior
- create a capability for iGEN to perform category learning.
An independent moderator benchmarked the improved COGNET/iGEN model against human performance data from a simplified Air Traffic Control task.
The COGNET/iGEN model came through with high marks. It accurately predicted human behavior for reaction times, error rates, and even subject self-assessments of workload under different Air Traffic Control conditions. Overall, COGNET/iGEN performed the best of all models tested (category learning comparisons are not yet complete). These new modeling features have been incorporated into the commercial iGEN human modeling software.Reference
Zachary, W., Santarelli, T., Ryder, J., Stokes, J., and Scolaro, D. (2001). Developing a Multi-tasking Cognitive Agent Using the COGNET/iGEN Integrative Architecture. In Proceedings of the 10th Conference on Computer-Generated Forces and Behavior Representation. IEEE/ITCMS, Piscataway, NJ, pp 79-90.
Zachary, W., Le Mentec, J-C, Iordanov, V. (2001). Generating Subjective Workload Self-Assessment from a Cognitive Model. In Proceedings of the Fourth International Conference on Cognitive Modeling, Fairfax, VA: Erlbaum.
Gluck, K., and Pew, R. (2001). Overview of the Agent-based Modeling and Behavior Representation (AMBR) Model Comparison Project. In Proceedings of the 10th Conference on Computer-Generated Forces and Behavior Representation. IEEE/ITCMS, Piscataway, NJ, pp 3-6.Tenney, Y., and Spector, S. (2001). Comparisons of HBR Models with Human-in-the-loop Performance in a Simplified Air Traffic Control Simulation with and without HLA Protocols: Task Simulation, Human Data and Results. In Proceedings of the 10th Conference on Computer-Generated Forces and Behavior Representation. IEEE/ITCMS, Piscataway, NJ, pp 15-26.
