Louis Nisiotis

1. The successful reintegration of offenders into society depends on their ability to acquire practical communication, decision-making, and social skills during incarceration. Traditional rehabilit...

Robotics and hyperautomation are moving from specialized tools to core capabilities for designing, operating, and governing complex systems. This article examines the convergence of robotics, digital twins, and artificial intelligence orchestration within cyber–physical–social systems, where robots act as embodied computing nodes across physical and digital worlds, highlighting enabling technologies and adoption pathways for scalable hyperautomation.

Quantised small language models (SLMs) deployed at the edge are increasingly important for enabling high performance AI agents in interactive virtual worlds, where low latency operation is key for seamless user experience. This paper investigates the deployment of quantised SLMs at the edge to meet performance requirements and maintaining high response quality to support real time conversational agent interactions in an immersive virtual world. We developed a RAG based back-end AI system to support virtual agent conversational capabilities in a virtual world prototype, and experimented with two SLM families and three quantisation levels (8-bit, 4-bit, and 3-bit) to evaluate system performance through end-to-end latency, time-to-first-token, throughput, and memory headroom. We also assess response quality via expert evaluations of accuracy, relevance, safety, and human alignment against non-quantised baselines. Results show that moderate quantisation enables performant practical edge deployment with good response quality, demonstrating that edge-deployed quantised SLMs can support real-time narrative interactions and remain practically useful for virtual-world agents, providing evidence for deploying high performance conversational AI in immersive environments.

The wireless communication industry and research community has focused on developing accurate and efficient models for signal propagation extensively over the past years. High-fidelity simulation software often comes with high costs and relies heavily on offline computations, limiting flexibility, and scalability. This article introduces a low-cost prototype for simulating signal propagation, leveraging advancements in gaming technologies, virtual reality (VR), and digital twins. The system is capable of conducting ray tracing simulations of wireless signal propagation through various configurable parameters and fidelity levels, generating signal strength heatmap and real-time ray visualization in immersive VR and 3-D desktop modes. Thorough device and performance testing validated the prototype’s development and assessed its implementation feasibility. The findings establish its operational capabilities and performance metrics, providing insights and guiding further development and refinements, ensuring its reliability and effectiveness in simulating complex wireless propagation.

The growing adoption of the Metaverse raises critical challenges in real-time crowd sensing, where traditional vision-based systems struggle to balance high-resolution monitoring with user privacy. Cameras and other optical sensors, while effective in tracking movement and interactions, inherently capture personally identifiable information, creating ethical and legal concerns. This paper explores the use of millimetre-wave (mmWave) radar technology as a privacy-preserving, high-resolution solution for real-time crowd sensing in Metaverse applications. Recognising the limitations of traditional monitoring methods, such as visual surveillance and mobile-based tracking, this study presents a simulation-based framework for evaluating mmWave-enabled visitor tracking within a museum-style environment. A MATLAB-based simulator models realistic human mobility and sensor data, incorporating error models obtained from an experimental precision analysis of mmWave sensors. A combination of DBSCAN and K-Means clustering is then applied to estimate crowd formations, density, and mobility flows. Results demonstrate the effectiveness of mmWave in identifying dynamic crowd behavior while preserving user anonymity, highlighting its potential for immersive digital twins, XR experiences, and intelligent environment management in Cyber-Physical-Social Systems that underpin the Metaverse.

VR gaming is rapidly growing in popularity, with puzzle-based games featuring optical illusions attracting gamers attention. Optical illusions are an emerging gameplay mechanic, challenging players' perceptual and cognitive abilities. While widely used in 2D/3D games, their integration into VR remains underexplored. This study investigates the use of optical illusion challenges in a VR game, focusing on players' workload demands, possible onset of VR sickness, and overall experience. A not-so-serious educational game featuring optical illusion challenges as core tasks for progression was developed for VR and PC, and a comparative evaluation was conducted. Results revealed that these challenges provided an enjoyable and immersive experience, with generally low workload demands. However, the complexity and disorienting nature of some illusions occasionally increased physical and mental effort, triggering VR sickness in some players. These findings offer valuable insights for designing VR games, highlighting the importance of balancing challenge and comfort when implementing such mechanics.

At the intersection of the physical and digital worlds, the metaverse is driving a new technological revolution across industrial applications, entertainment and gaming, transportation, and numerous other sectors.

This paper proposes a system architecture for deploying AI-driven interactive systems within the context of a Cyber-Physical-Social System framework, specifically designed for applications in the emerging Metaverse, applicable to a plethora of domains. The proposed architecture integrates advanced technologies to facilitate immersive user experiences and intelligent system interactions based on user actions and realtime data inputs from the real and the digital world(s), digital twins, artificial agents and elements. The integration of the proposed architecture components aim to support sophisticated interactions between physical, digital entities, and humans within the Metaverse. The paper presents the theoretical underpinnings and system description of the proposed architecture, contributing to future developments in Metaverse systems and technologies. This article has been accepted for publication in IEEE International Symposium on Emerging Metaverse (ISEMV 2024) . This is the author's version which has not been fully edited and content may change prior to final publication. Citation information : L. Nisiotis, "A System Architecture for AI-Driven Interactive Systems Within a CPSS Framework for Metaverse Applications,"  2024 IEEE International Symposium on Emerging Metaverse (ISEMV) , Bellevue, WA, USA, 2024, pp. 9-12, doi : 10.1109/ISEMV63338.2024.00016.  

As educational technology evolves, the potential of Pedagogical Agents (PAs) in supporting education is extensively explored. Typically, research on PAs has primarily focused on computer-based learning environments, but their use in VR-based environments and integration into education is still in its infancy. To address this gap, this paper presents a mixed method comparative study that has been conducted to evaluate and examine how these computer-based PAs and VR-based PAs compare, towards their learning efficacy and technology acceptance. 92 Computing and Engineering undergraduate students were recruited and participated in an educational experience focusing on computing machinery education. The findings of this study revealed that both approaches can effectively facilitate learning acquisition, and both technologies have been positively perceived by participants toward acceptance, without any significant differences. The findings of this study shed light on the potential of utilizing intelligent PAs to support education, contributing towards the advancement of our understanding of how to integrate such technologies to develop learning interventions, and establishing the foundation for future investigations that aim to successfully integrate and use PAs in education.