Human Performance Models : applications to complex and operational tasks – MPH

With scientific and technological advances, our world is becoming more complex. In particular, operators, who benefit from these advances, are faced with increasingly sophisticated tasks. Despite advances in automation and artificial intelligence, humans still play an essential role in managing and supervising these complex systems. Although they remain more flexible than automated systems, they are still susceptible to errors. Statistics in aeronautics show, for instance, that human factors are at the origin of most serious accidents. A lever for action could be the creation of human performance models to understand how these errors occur and thus prevent them. Research in neuroergonomics has allowed the identification of neurophysiological markers of fatigue, mental load or attentional tunneling, which are factors that reduce performance (Dehais et al., 2020) or, on the contrary, markers of cognitive efficiency (Chenot et al., 2021). However, this research has most often focused on a single parameter of human functioning and via the prism of a single method of investigation (electrophysiology, electrocardiography, behavior, etc.). This results in the absence of a holistic approach to the understanding of human physiology and cognition in the context of complex tasks.
The MPH project therefore aims at developing and validating in an operational situation a complete psychophysiological model of human performance during complex tasks, applicable to civilian or military people in learning situations (e.g., students, engineers, medical interns, etc.) or in complex systems management situations (e.g., airplane or drone pilots) through 3 objectives :
1. Predict human performance from intrinsic measures. The aim is to predict performance during the realization of complex tasks, and in particular errors, from intrinsic psychophysiological measures. Correlations will be made between cognitive/physiological functioning (cognitive architecture and/or functional brain connectivity) measured at rest and performance on complex tasks.
2. Build a mental state estimation tool. Norms will be established on a large sample of individuals who will be subjected to a battery of complex tasks and physiological measures. All these measures will allow a complete and robust modeling of human performance. Machine learning algorithms based on physiological signals (brain, heart and eye activity) will be developed to estimate mental state during complex tasks.
3. Operational validation. The validation of this tool will be done by testing it on new complex tasks (i.e., transfer) in operational situations during ecological experiments (drone simulators with military and civilian pilots and aircraft simulators with civilians).
The MPH project could have a direct impact on :
- Academic and fundamental research, by providing a qualitative, quantitative and perennial database, which can serve as a foundation for numerous research projects on the links between physiological activity and cognitive performance.
- Applied research (in particular in aeronautics) by providing a model of human performance usable in ecological situations (TRL level 4 to 5), with the aim of improving safety in risky systems.