Influence of Automatic Speech Recognition and Understanding on Flight Efficiency and Throughput - A Human-in-the-Loop Study

Modern Controller Working Positions (CWP) require manual input into digital flight strips or aircraft radar labels. Erroneous or missing input from air traffic controllers (ATCos) can lead to inefficiencies and potential safety issues. Automatic speech recognition and understanding (ASRU) support is a method to improve the accuracy of digital system inputs. This paper investigates the effectiveness of ASRU support on ATCos radar label maintenance, with the aim of determining the impact on of Air Traffic Management efficiency. To that aim, a human-in- the-loop study was carried out at DLR Braunschweig. The results reveal (i) that the verbal ATCo transmissions analyzed by ASRU achieved a robust command recognition rate of 92.5% with a command error rate of 2.4%; (ii) within the baseline conditions seven actual safety-critical situations were observed, whereas with ASRU support only four happened; (iii) through ASRU support the number of wrong or missing radar label entries were reduced by the factor of two; and (iv) ATCos' situational awareness improved while the perceived mental workload decreased. These results indicate that more mental spare capacity was available with ASRU support, which improves the planning process and increases traffic flow.