How can we measure trust, workload, and awareness in real-time?

Imagine working in a team with a robot in space. How do you know if you're managing stress, trust, and situational awareness well? A new study has developed psychophysiological models that can measure these aspects in real-time, without needing specific information about each operator. Ten participants monitored a simulated space habitat, and researchers used regression models to predict levels of trust, workload, and awareness. The results show that these models work well, with an accuracy of 67% for trust, 60% for workload, and 85% for awareness. This could revolutionize the way we work with robots, making our teams more efficient and safe.

The psychophysiological models developed in this study use regression techniques with LASSO-based feature selection. This method helps identify the most relevant variables for predicting levels of trust, workload, and awareness, improving model stability and generalizability. Leave-one-participant-out cross-validation was used to assess the predictive ability of the models, demonstrating that they can be applied to new operators without specific information. This approach could be extended to other contexts, such as autonomous driving or healthcare, where understanding the cognitive state of operators is crucial.