This report discusses how the Boeing AugCog Team approached Augmented Cognition (AugCog) in a closed-loop simulation of an unmanned air vehicle (UAV) system. The Boeing AugCog system was comprised of the UAV simulation and interface, a sensor suite of multiple physiological sensors of the operator's cognitive state, cognitive state assessor (using an artificial neural net and Statistical Process Control software), data collection software called CWAD (TM), and an augmentation manager which dynamically alters the user interface in response to changes in the UAV operator's cognitive state. Phase 2 consisted of alternating periods of technology development and technology assessment. Development of specific system components was conducted remotely at each team member site. This presented a number of challenging problems for the Boeing Team. For example, Boeing was responsible for development of the augmentation manager and cognitive bottleneck mitigations, but did not have the sensors and cognitive state classifiers to test the effectiveness of the mitigation concepts. To work around this problem, we tested the mitigation software by assuming the presence of a perfect sensor that always detected cognitive overload at an appropriate point in the mission scenario. Pilot studies were then conducted to assess operator performance under mitigated and non-mitigated conditions. The mitigation concepts were iteratively refined until the desired level of performance enhancement was achieved. Concept validation experiments (CVE) were conducted bi-monthly to test the effectiveness of the mitigation concepts for each cognitive bottleneck in a closed-loop setting. All CVEs were conducted at the Integrated Technology Development Laboratory in Seattle, Washington. Separate experiments were constructed and conducted for each of five cognitive information processing bottlenecks: 1. Working Memory, 2. Executive Function (our assigned primary bottleneck), 3. Sensory Input, 4. Attention, and 5. Response Generation. During experimental trials, the Boeing AugCog team collected and analyzed data for aided (closed-loop) and un-aided conditions with four trained UAV operators. The results of experimental trials demonstrate that aiding from closed-loop augmented cognition improves UAV operator performance significantly.
