Integrated Human Behavior Modeling and Stochastic Control (IHBMSC)

• Main Authors: Chandler, Phillip R, Patzek, Michael J, Pachter, Meir, Rothwell, Clayton, Naderer, Sara, Kalyanam, Krishna
• Format: Report
• Language: English
• Subjects: ALGORITHMS; ALTITUDE; ASPECT ANGLE; ATR(AUTOMATIC TARGET RECOGNITION); ATR-HUMAN TEAM; AUTOMATION; BEHAVIOR; CONFUSION MATRIX; COOPERATIVE CONTROL ALGORITHMS; CUES(STIMULI); DETECTION; DWELL TIME; DYNAMIC PROGRAMMING; DYNAMIC PROGRAMMING ALGORITHMS; ERRORS; FA(FALSE ALARMS); FALSE ALARMS; FUNCTIONS(MATHEMATICS); Human Factors Engineering & Man Machine System; IDIOGRAPHICS; IG(INFORMATION GAIN); IHBMSC(INTEGRATED HUMAN BEHAVIOR MODELING AND STOCHASTIC CONTROL); LEAP II; LEAP(LEARNING ESTIMATES OF AGGREGATE PERFORMANCE); MAN MACHINE SYSTEMS; MATRICES(MATHEMATICS); MD(MISSED DETECTION); METRICS; Numerical Mathematics; OBJECTIVE FUNCTION; OPERATOR CUEING; OPERATOR ERROR MODELING; OPERATORS(PERSONNEL); PE62201F; PERFORMANCE METRICS; RELIABILITY; REMOTELY PILOTED VEHICLES; SCENARIOS; SDT(SIGNAL DETECTION THEORY); SEQUENTIAL INSPECTION; Statistics and Probability; STOCHASTIC CONTROL; TARGET DENSITY; Target Direction, Range and Position Finding; TARGET DISCRIMINATION; TARGET RECOGNITION; TEST AND EVALUATION; UAV(UNMANNED AERIAL VEHICLES); UNCERTAINTY; UNMANNED; VSCS GROUND STATIONS; VSCS(VIGILANT SPIRIT CONTROL STATION); WORKLOADS; WUAFRL2403Q05C

This project developed an operator error model for improving the accuracy of the operator task of discriminating objects in multiple UAV videos on the Vigilant Spirit Control Station (VSCS) ground station. This error model is idiographic in the sense that it is tailored for a specific individual. The operator error model is used in a stochastic controller which decides, based on the model uncertainty and the operator input of target/non-target, whether to revisit the object to obtain additional information or to continue to the next object. The stochastic controller computes the probability of a target given the available evidence and the reliability of the evidence as per the operator error model. Over 120 operator-in-the-loop tests were conducted to obtain sufficient data to be able to estimate the false alarm (FA) and missed detection (MD) rates over four variables (altitude, aspect angle, dwell time, workload). For a 50% target density scenario where the operator discrimination task is fairly difficult, the integrated human behavior model and stochastic controller (IHBMSC) was shown to reduce the FA rate from 35% to 5%. Also, the IHBMSC performance was found to be quite robust for lower target densities (5%) and different operators. Prepared in collaboration with the Supervisory Control and Cognition Branch, Warfighter Interface Division, Air Force Research Laboratory, 711 Human Performance Wing, Wright-Patterson Air Force Base, OH, the Air Force Institute of Technology, Wright-Patterson Air Force Base, OH, and InfoSciTex Corporation, Dayton, OH.