Back to search

National Technical University of Athens (NTUA)

Συνεργασία ανθρώπου-ρομπότ: ενίσχυση της προβλεπτίκης συμπερίφοράς του άνθρώπου

Abstract

dc:description

This doctoral dissertation provides a systematic investigation of Human–Robot Collaboration (HRC), focusing on one of its most fundamental yet insufficiently explored aspects, i.e. the enhancement of human anticipatory capability. In today’s landscape, human–machine interaction is shifting from supervisory roles to genuine forms of joint task execution. An increasing number of collaborative robots are being deployed in industry, working alongside humans to complete shared tasks. The thesis develops a comprehensive theoretical framework in which human–robot collaboration is approached as a form of joint action analogous to human–human cooperation. Within this framework, the human ability to anticipate a partner’s actions and intentions is a critical prerequisite for safe and smoothly coordinated collaboration. Correspondingly, the robot’s capacity to render its upcoming actions legible and easily interpretable is essential for building trust, preventing collisions, and fostering the development of a compatible mental model of action. To empirically investigate these theoretical premises, two high-fidelity Virtual Reality (VR) experimental environments were developed. These allowed for precise control over collaboration parameters and safe simulation of complex industrial scenarios. The impact of anticipatory visual cues— color-coded signals revealing the robot’s imminent movements—was evaluated across a set of operational and cognitive metrics, including task completion time, number of collisions, synchronization ability, perceived safety, and participants’ cognitive strategies. The first chapter begins with the definition of the robot and the historical evolution of Robotics, from ancient automata to modern intelligent and collaborative robots. It presents the theoretical components of Human–Machine Interaction (HMI), such as cognitive models, usability, and function allocation, which determine how users interact with technological systems. At the same time, contemporary technological developments (sensors, artificial intelligence, connectivity, advanced materials, wearable devices) and the framework of safety standards are examined, emphasizing that effective and safe collaboration does not depend solely on technical performance, but primarily on the degree of cognitive coupling, transparency, and predictability achieved in dynamic environments. The second chapter examines the transition from classical Human–Robot Interaction (HRI) to modern Human–Robot Collaboration (HRC). Initially, Human–Robot Interaction and its unique characteristics are analyzed, especially in relation to robot autonomy, perception, and adaptability. Key taxonomies of interaction and the possible roles assumed by humans are presented, along with the communication modalities used between the two partners. The chapter concludes with the concept of Human–Robot Collaboration, highlighting its benefits in terms of productivity, safety, and workflow efficiency. In addition, the categories and evaluation metrics used to assess the quality of collaboration are presented, contributing to the design of more effective and human-centered collaborative systems. The third chapter explores the role of anticipatory and adaptive behavior in Human–Robot Collaboration, showing how these mechanisms enhance coordination, safety, and naturalness of interaction. Anticipation enables prediction of forthcoming actions, while adaptability supports flexible responses to changing conditions. The chapter outlines the benefits of these mechanisms and the cues used for intention prediction. Speech Act Theory is introduced as a unified lens for interpreting human–robot communication, and the analysis extends to autonomous vehicles, where prediction and behavioral transparency are likewise crucial. The chapter concludes with a literature review of approaches used by robots to infer human intentions and of the ways humans interpret robotic behavior, emphasizing the importance of predictability and transparency for effective collaboration. The fourth chapter presents the research questions of the thesis and the two experimental studies conducted to evaluate the impact of anticipatory and adaptive mechanisms on human–robot collaboration. Both studies were implemented in Virtual Reality, enabling precise control of the collaborative environment and safe investigation of human behavior. The results demonstrated that anticipatory cues exert a strong and immediately beneficial influence on collaboration. Their presence led to faster task completion, substantial reduction in collisions, improved synchronization, smoother interaction, and reduced levels of hesitation and uncertainty. Subjective assessments further supported these findings, indicating that visual anticipatory cues enhance the transparency of robot behavior, increase environmental predictability, and strengthen participants’ trust. Thus, the improvement concerns not only operational efficiency but also acceptance of robots and willingness to collaborate. At the same time, adaptive robot slowing clearly improved safety by reducing collision likelihood. However, this strategy was not accompanied by equivalent improvements in efficiency or collaborative fluency, suggesting that passive adaptation mechanisms, while vital for user protection, are insufficient on their own to achieve high-quality collaborative behavior. The findings also showed that predictive cues become even more critical under high-speed conditions, where synchronization demands rise and agents require faster, more accurate understanding of each other’s intentions. Moreover, prior exposure to predictive cues led to sustained performance even after their removal, indicating learning effects and the formation of mental models of robot behavior. Yet, in some cases, overconfidence emerged, increasing collisions once cues were withdrawn, highlighting the need for careful cue design. Finally, comparing virtual and real-world conditions revealed that key patterns of human behavior—such as risk assessment, synchronization, and collision avoidance—remain remarkably consistent, reinforcing the external validity of the methodology and confirming the usefulness of virtual environments for designing and evaluating collaborative systems.Beyond its empirical contributions, the thesis offers significant methodological and practical value. The developed experimental platform can be used to systematically study the difficulty of collaborative tasks, evaluate different collaboration schemes, formulate design guidelines, and extract optimization rules that balance efficiency and safety. Furthermore, novel evaluation approaches are proposed, applicable both for comparing different interfaces and robotic behaviors and for informing industrial standards. Overall, the thesis concludes that deliberate communication of robotic intentions serves as a catalyst for a new era of human–machine collaboration. Such collaboration relies not only on technical proficiency but on the cognitive, psychological, and functional synergy that emerges when robot actions become legible and predictable. In a rapidly evolving industrial landscape where humans and robots are expected to coexist, this thesis provides a systematic framework for the design of human-centered collaborative systems.

Degree

thesis:*
Grantor dc:publisher
National Technical University of Athens (NTUA)
Year dc:date
2026

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Psarakis, Loizos
  • Ψαράκης, Λοϊζος

Subjects

dc:subject × 16

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/62366

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Psarakis, Loizos; Ψαράκης, Λοϊζος. Συνεργασία ανθρώπου-ρομπότ: ενίσχυση της προβλεπτίκης συμπερίφοράς του άνθρώπου. National Technical University of Athens (NTUA), 2026. http://hdl.handle.net/10442/hedi/62366