UC Berkeley

Communication between pilots and controllers plays an important role in the Air Traffic Control (ATC) process. There have been many studies on the ATC communication analysis but rarely on the suppression of voice activity. A censored regression model with a stochastic right-censored threshold is developed to estimate the relationship between ATC communication voice activities and the ATC operational environment through a wide and representative set of real operational ATC voice recordings. The model estimation result shows that flight operational volume, lead flight operations, time of day, meteorological conditions, visibilities, wind conditions, anomalous flight operations and runway configurations are good indicators of the ATC communication. Lastly, we use the estimated model to simulate the loss percentage of ATC communication. This study has profound implications on further understanding of ATC system capacity, aviation safety and human performance interfacing ATC.