Establishing trust in decision-aiding: The role of a visualization of agent decision-space in improving operator engagement with reduced cognitive load

Decision recommendation systems relieve operators from high cognitive-load during stressful situations. However, automation over-trust can disengage complacent operators from the task leading to lower situation awareness and inability to intervene and override incorrect recommendations. Our recent research effort was focused on a visualization of agent decision-space to improve automation transparency and aid the operator's perception of the environment. We describe specific properties of the interface and their anticipated benefits such as improved situation-awareness and expectancy. The visualization is compared with an alternative static representation with an emphasis on how the visualization improves expectancy. An experiment was conducted with a command and control simulation environment to compare the two representations. The results of the experiment have been encouraging. Observed performance improvements in specific scenario conditions, are in accordance to anticipated benefits of the visualization..