Real-time Continuous Uncertainty Annotation (RCUA) for Spatial Navigation Studies

Navigation uncertainty is crucial for understanding the wayfinding behaviors while no method has been developed to effectively measured uncertainty in real-world scenarios. We developed the Real-time Continous Uncertainty Annotation (RCUA) to continuously measure perceived uncertainty by asking users to push the joystick in during the wayfinding process. We tested its test-retest reliability and validated RCUA based on 40 participants using the known group and known treatment. We also compared it with a discrete self-report scale and continuous postexperiment video annotation (CUA). The result demonstrated that most participants were able to output four distinct levels of uncertainty, though high variability and errors were observed. Both known group and known treatment proved good validity of the measure and RCUA was moderately correlated with self-report uncertainty. Self-report surveys showed that participants can continuously push the joystick and conduct wayfinding tasks at the same time. A comparison between RCUA and CUA showed that RCUA had a higher granularity but participants tended to overreport uncertainty using RCUA.