Towards the Wearable Cardiorespiratory Sensors for Aerospace Applications

In safety-critical aviation operations, adaptive Human-Machine Interfaces (HMI) rely on accurate physiological monitoring to mitigate cognitive overload. While cardiorespiratory sensors are promising for real-time cognitive workload assessment, existing studies lack rigorous validation of consumer-grade devices in high-stress aviation contexts and fail to address measurement uncertainty propagation. This study evaluates the Zephyr BioHarness, a commercial wearable sensor, against clinical-grade equipment during arithmetic tasks simulating aviation cognitive demands. By integrating a neuro-fuzzy system with uncertainty propagation methods, we quantify the reliability of heart rate (HR) and breathing rate (BR) metrics for workload estimation. Results demonstrate moderate HR accuracy (RMSE: 4.85 bpm, CC: 0.66) but poor BR performance (RMSE: 9.73 bpm, CC: 0.09), attributed to inconsistent breath detection during cognitive strain. The novel uncertainty framework reveals workload prediction variances (σWL: 0.38–2.22) driven primarily by BR inaccuracies, emphasizing the need for improved respiratory sensing in adaptive HMI. This work pioneers the application of neuro-fuzzy systems for uncertainty analysis in aviation physiology, offering a validated methodology for sensor integration and highlighting critical limitations in current consumer-grade technologies. These findings advance the design of robust cognitive monitoring systems, ensuring safer and more efficient human-machine collaboration in aviation.