T22

TX22 dataset: Predicting and influencing trust-based decisions about control authority hand-off and take-over during simulated, semi-automated driving in a leader-follower paradigm.Vehicle survivability is critically important in todays military. Significant DoD investments have focused on developing and integrating autonomous vehicle technologies to mitigate the effects of human error and thus enhance surviability and mission effectiveness. In a previous experiment (SANDR designation: ARL_TX20), we explored how a human operators acceptance and use of advanced technology is influenced by their trust and related factors, like subjective workload and automation reliability. Nevertheless, more critical than measuring and achieving a certain level of trust is the need for a capability to resolve observed (or predicted) discrepancies between trust and trustworthiness that will undermine effective joint system performance. Using the same paradigm as we developed for our previous experiment (ARL_TX20), here we explore our ability to (a) make accurate real-time predictions of instances where intervention is necessary and (b) use those predictions to provide feedback to the driver that is intended to support active "trust management" by influencing the trust-based decisions of the driver.