Cognitive Augmentation Technologies: VR, AI and Social Robots for Industry 5.0

Licentiatavhandling, 2025

Digital innovations and an increasingly integrated global economy are transforming manufacturing systems—complex sociotechnical arrangements that integrate technology, information systems, and human labor. For European industry in particular, the COVID-19 pandemic, supply chain disruptions, and trade tensions have heightened uncertainty and exposed vulnerabilities in conventional production models. Besides, manufacturing is undergoing rapid transformation amid an aging workforce, prompting a shift toward Industry 5.0’s human-centric vision. This paradigm seeks to harmonize advanced technologies with inclusive design, worker empowerment, and well-being, addressing widening skill gaps and the need for resilient, sustainable operations. Within this context, cognitive augmentation technologies—such as video-based instruction, virtual reality (VR), and AI-enabled humanoid social robots—offer new pathways to support learning and day-to-day operations, but also introduce mental workload demands that must be carefully measured and managed.

Comparative studies show that augmentation yields the greatest value when systems are designed for accessibility, usability, and well-being. Video instruction proved most cognitively efficient for onboarding; VR delivered superior spatial understanding but imposed the highest mental workload; AI-driven humanoid social robots occupied a middle ground. Overreliance on AI guidance increased mental workload and error risk, underscoring the need for calibrated assistance, transparent interaction, and thoughtful human–AI task allocation. Human-centered interface design, clear information pacing, and adaptive guidance emerged as key levers to reduce unnecessary complexity, support universal design demands, and enhance satisfaction and motivation.

The thesis demonstrates that a mixed-measures approach—integrating subjective ratings (e.g., NASA-TLX, RSME), physiological signals (e.g., HRV), and performance indicators (completion time, error rate)—provides a robust basis for quantifying mental workload and comparing modalities. Triangulation enabled detection of overload and underload, informed iterative interface refinements, and supported closed-loop adaptations in future research (e.g., tuning information density and pacing, simplifying interaction flows, switching assistance modality) to maintain workload within an optimal band.

Conceptually, the work reframes Industry 5.0 as a human-centric ecosystem where technologies symbiotically augment and empower workers, supporting dignity, wellbeing, and inclusion. Methodologically, it contributes a validated pipeline for mental workload assessment in real manufacturing contexts based operations. Practically, it offers guidance for technology development and adoption in adaptive and inclusive systems for universal design for vulnerable groups, which is also the future direction to continue after the licentiate study. Together, these contributions lay a foundation for resilient, inclusive, and sustainable manufacturing aligned with the goals of Industry 5.0.