Transparent Tracker

The goal of this research is to understand how to structure an environment in which RPA Sensor Operators can manage tracking multiple targets with an automatic tracker rather than tracking one target track at a time. Our approach applies Chen’s Situation-awareness based agent transparency model: An object-tracking algorithm will report its SA to Sensor Operators (SO). SO will use the SAT level to assist them in managing automatically tracked targets. . We will begin the work by conducting a Cognitive Task Analysis of SO job. Results will inform the design of interface elements that convey the algorithm’s SAT to SO. We will modify an existing excellent object tracker to report its SAT values. SAT values capture the current state of the algorithm, factors affecting its reasoning, and predictions. We will design interface elements that communicate the SAT values from the algorithm to SO. Then we will design an environment in which the effectiveness of delivering SAT information to users can be evaluated. Experiments will be conducted to measure how well users SO use the information presented. We will demonstrate this environment to sponsors and discuss future project direction.