Lo Regio, Fabrizio (2026) A metrological framework for XR-based natural human-machine interfaces: interaction engineering in Industry 5.0. [Tesi di dottorato]

Documento PDF LoRegio_Fabrizio_38.pdf Visibile a [TBR] Amministratori dell'archivio Download (44MB) | Richiedi una copia

Abstract

The paradigm shift toward Industry 5.0 redefines the industrial landscape by placing the human operator at its core, necessitating a symbiotic integration between physical actions and digital systems. Within this framework, eXtended Reality (XR) has evolved from a simple visualization technology into a sophisticated Human-Edge platform where the XR Head-Mounted Display functions as a critical terminal node. This allows Natural Human-Machine Interfaces (NHMI) to interpret innate human behaviors, ranging from gaze and gestures to brain activity, and convert them into digital commands. However, the effective integration of these interfaces into safety-critical industrial and healthcare domains is currently obstructed by a significant Metrology Gap. While future networks promise highly reliable communications, the reliability of the entire chain is compromised at the source if the generation of the data itself is not trustworthy. Current evaluations predominantly rely on qualitative usability studies, which fail to provide the objective guarantees required for operational trustworthiness or to address the opacity of commercial hardware specifications. This gap is characterized by the absence of standardized characterization protocols, the opacity of commercial hardware specifications, and, critically, the lack of rigorous methods to quantify the stochastic impact of human variability on system performance. To bridge this gap, this doctoral thesis establishes the discipline of Interaction Engineering, shifting the paradigm from subjective assessment to rigorous metrological characterization. By treating the NHMI not as a consumer gadget but as a measuring instrument, this thesis proposes a unified framework compliant with Guide to the Expression of Uncertainty in Measurement (GUM). This framework characterizes a wide spectrum of HMIs. The scientific contributions of this work include the development of a standardized modular testbed and the formulation of novel figures of merit to benchmark device capacity independently of hardware specifications. A pivotal innovation of this methodology is the formulation of a comprehensive measurement uncertainty budget that explicitly decomposes performance variability into three distinct categories: instrumental uncertainty, intra-individual variability associated with user repeatability, and inter-individual variability reflecting reproducibility across the user cohort. Experimental validation, conducted using the Microsoft HoloLens 2 as a representative case study, demonstrated the framework’s capability to rigorously discriminate performance across diverse interaction paradigms and reveal critical trade-offs between information throughput and spatial precision. The proposed methodology successfully quantified distinct operational profiles, revealing critical trade-offs between performance, while effectively addressing the evaluation of the variability introduced by the users. Ultimately, this work validates the proposed framework as a device-agnostic infrastructure essential for certifying NHMIs as reliable measurement instruments, thereby ensuring the quality of Human-Edge data required for the Industry 5.0 ecosystem.

Downloads

Actions (login required)

Modifica documento