This thesis focuses on detecting variations in cognitive state in demanding professional situations using non-invasive physiological measurements. It proposes a systemic approach combining the analysis of motor and cardiac dynamics, made possible by the development of a multimodal cyberphysical system (CPS). Three experimental protocols identified relevant indicators for monitoring cognitive state: during a task of increasing difficulty, after a heart coherence session, and under typical working conditions. The results highlight the sensitivity of multifractality and non-linearity indices of movement, as well as cardiac markers in the frequency domain, to assess variations in cognitive state. Finally, the real-time integration of cardiac analyses within the CPS paves the way for physiological feedback that can improve comfort, safety, and performance in real-world situations.
