Human-Machine Performance Envelope: Controller Adaptive Digital Assistant evaluation

Air traffic control operators face sustained cognitive demands that can elevate stress, induce fatigue, and impair vigilance. Integrating assistant tools can help alleviate rising workload levels, enabling air traffic control operators to manage increasing traffic volumes while maintaining operational performance. This work employs the Human–Machine Performance Envelope framework designed to capture and assess the dynamic interaction between air traffic control operators and support artificial intelligence systems. To demonstrate its applicability, the Human-Machine Performance Envelope is applied to the CODA (COntroller adaptive Digital Assistant) system, a prototype digital assistant developed to proactively support the activity of the air traffic controllers and reduce mental workload. CODA operates through three functional stages—real-time data ingestion and short-term prediction, context-aware automation adaptation, and dynamic flight control allocation—facilitating adaptive sharing of responsibilities between human operators and automation. Using the Human- Machine Performance Envelope framework, we evaluate the integrated solution’s impact on key cognitive and operational metrics, including stress, fatigue, vigilance, and human–AI coordination. Results indicate that the Human-Machine Performance Envelope reliably captures both objective and subjective aspects of controller–AI interaction.