Decision Modeling
Program Description
As the exposure of autonomous systems increases within the various services in society, it has become increasingly apparent that there is a need for understanding how humans interact with these systems on the user decision level. A salient question is: How best should the human element be integrated within such a system? It is clear that too much input floods the user, such that many user decisions are rushed. It is also clear that too little input bores the user, such that reaction ability and time can be suspect. It is also well known that humans are very good at complex strategies, while machines are very good at repeating fast, well known commands. While it appears that the “best” interface between a user and the system is between these extremes, it is also noted that user decisions (type and frequency) may change as the environment changes.
The objective of this program is to develop and validate modeling methodologies which can encapsulate human user decision making when controlling complex systems involving many autonomous agents. Specifically, this program develops hybrid-stochastic system identification techniques that can be used to identify a hybrid model of user decisions from user data when controlling a complex system. This model will enable the probabilistic evaluation of an “average” and a “deviation” of each decision, as well as their dependency on environmental parameters. It is envisioned that the identified hybrid model can be used to evaluate when users have difficulty making decisions. The AFRL Human Effectiveness Directorate will lead a series of experiments with AFRL subject pool to develop user data for identification and evaluation; the end experiment will evaluate adaptive tasking, an approach where decision number/complexity adapt based on the changing environment.
Sponsor
AFOSR
Staff and Students
- David Schneider, ME Ph.D., Student Teaming Approaches for Multiple Vehicle Coordination and Control
- Jesse Veverka, AE Ph.D., Student Operator Decision Modeling in Semi-Autonomous Multiple Vehicle Systems
Publications
- D. Schneider, M. Campbell, “Real Time Optimal Task Allocation in Highly Dynamic Environments,” submitted to the ASME Journal of Dynamic Systems, Measurement and Control.
- J. Veverka, M. Campbell, “Operator Decision Modeling for ISR Type Missions,” IEEE Conference on Systems, Man, and Cybernetics, Oct 2005.
- D. Schneider, M. Campbell, “Real Time Optimal Task Allocation in Highly Dynamic Environments,” ASME International Mechanical Engineering Congress and Exposition, Nov 2005.
- J. Veverka, M. Campbell, “Towards an Operator Decision Model for ISR Type Missions,” AIAA Guidance, Navigation and Control Conference, Aug 2005.
- M. Campbell, R. D’Andrea, J.-W. Lee, and E. Scholte, “Experimental Demonstrations of Semi-Autonomous Control,” 2004 American Control Conference, June 2004.
- Veverka, J. and Campbell, M., “Experimental Study of Information Load on Operators in Semi-Autonomous Systems,” 2003 AIAA Guidance, Navigation and Control Conference, Austin TX, Aug. 2003.
- J. Sullivan, S. Waydo, and Campbell, M., “Using Stream Functions to Generate Complex Behavior,” 2003 AIAA Guidance, Navigation and Control Conference, Austin TX, Aug. 2003.
- A. I. Chaudhry, R. D’Andrea, and M. Campbell, “RoboFlag – A Framework for Exploring Control, Planning, and Human Interface Issues Related to Coordinating Multiple Robots in a Realtime Dynamic Environment,” 11th International Conference on Advanced Robotic, Portugal, June 2003.
- Campbell, M., D’Andrea, R., Schneider, D., Chaudhry, A., Waydo, S., Sullivan, J., Veverka, J., Klochko, A., “RoboFlag Games using Systems Based, Hierarchical Control,” American Control Conference, June 2003.
