Theoretical and Methodological Basis of the Concept of Perceptual-Cognitive Interfaces

The authors propose a concept of perceptual-cognitive interfaces (PCIs) for modern complex ergatic technical systems (MCETSs), which are dynamic cognitive ecosystems comprising humans, AI agents, sensors, and robots. Traditional interfaces fail to cope with uncertainty and the high speed of decision-making. The PCI concept addresses this problem by introducing an autopoietic cognitive partner embedded in the perception-action-reflection loop. PCIs amplify human capabilities through: 1) Bayesian predictive models of the activity world; 2) multimodal data fusion (video, biosignals, eye-tracking, tactile signals); 3) synergistic real-time interaction; 4) formation of emergent affordances. The theoretical basis of the PCI concept is the enactive approach, predictive processing, autopoiesis, and extended cognition. The following postulates are formulated: embodiment of cognition, predictivity of perception, enactive interaction, affordance specificity, cognitive optimality, social enaction, dynamic adaptability, and ecological validity. By synthesizing these theories, PCIs minimize the operator’s cognitive load (internal/external/productive), construct dynamic knowledge network graphs, and support multi-agent for collective cognition. The methodology includes multimodal integration and human-centred design, consisting of four phases: enactive analysis, predictive modelling, affordance-based design, and adaptive implementation. Effectiveness is evaluated according to the criteria of cognitive efficiency, enactive naturalness, predictive coherence, adaptive flexibility, social support, and emergent synergy. The PCI concept sets a paradigm for the transition from local interfaces to distributed cognitive ecosystems, shifting the focus from information transfer to creating a shared space for cognitive activity.