Advances in artificial intelligence and embodied interaction in augmented reality (AR) are creating new opportunities for intelligent instructional systems that dynamically adapt to individual learners. However, sustaining real-time responsiveness while preserving natural, socially meaningful interaction remains a persistent challenge. This work introduces ARIA (Augmented Reality Instructional Agent), a real-time software architecture for embodied AI instruction in augmented reality. ARIA leverages large language models (LLMs) with adaptive prompt engineering to tailor dialogue style, instructional strategy, and persona expression to each user. Its modular pipeline is optimized for robust, low-latency performance, with benchmarks reported for responsiveness and system stability. To complement the technical evaluation, user experience was assessed through standardized questionnaires, offering insights into perceived personalization, trust, and interaction quality. Quantitative and qualitative results demonstrate that ARIA achieves sub-second responsiveness, high pragmatic and hedonic usability, and a strong sense of co-presence and instructional trust. This work contributes a unified framework and reference architecture for developing adaptive embodied agents that combine technical efficiency with human-centered design, highlighting how real-time responsiveness can serve as the foundation for relational engagement in embodied AI instruction.

While embodied conversational agents driven by Large Language Models (LLMs) are emerging as valuable tools for instruction in Augmented Reality (AR), a key challenge lies in crafting their personalities to optimize both instructional efficacy and user engagement. To address this, we present findings from a within-subjects experiment that compared task performance and user experience with a LEGO assembly task. Participants received guidance from a real human instructor and three virtual counterparts, whose Big Five personality profiles were designed to be: (1) a direct replica of the real human, (2) an "ideal" profile ba sed on pedagogical research, or (3) customized by the participant. Our results reveal a critical trade-off: instruction from the real expert resulted in superior task efficiency and clarity; however, among the virtual conditions, instructors with idealized or user-customized personalities fostered significantly higher levels of user engagement and social presence compared to the virtual replica. Crucially, allowing users to customize their instructor's persona led to the strongest preference for future interaction. These findings underscore that personality is a fundamental component in the design of AI-driven instructors, providing empirical evidence for navigating the balance between task-oriented guidance and personalized, socially resonant user experiences.

Teleportation has become a popular locomotion method for virtual reality due to lesser demands on physical space and decreased levels of motion sickness compared to other methods. However, prior work has shown that these advantages come at the cost of impaired spatial perception and awareness, the extent to which is still largely unknown. In this work, we present a within-subjects study (N = 29) that explores the effects of teleportation on spatial perception by investigating how much humans can be unknowingly displaced relative to their intended destination during teleportation. After teleporting to the specified location, participants indicated the direction and magnitude (small, medium, large) of the perceived shift or rotation. Displacement from the target happened either as a translation in the forward- or strafe-axis, or a rotation about the up-axis at the intended target. Each displacement condition included eleven offsets that were repeated six times. Our results indicate points of subjective equality, which show a significant perceptual shift along the forward-direction, as well as detection thresholds, which indicate a comparatively wide range in which humans are unable to detect induced shifts. Furthermore, our results show that even if humans are able to detect these shifts, larger ones can be introduced before their magnitudes are rated as medium or large, which provides ample opportunities for interface designers who want to leverage these results in virtual reality.

The swift advancement of drone technology presents new challenges in data analysis and integration as deployments increase. This paper addresses the complexities of managing multiple drone video feeds and controlling autonomous drone swarms to enhance situational awareness through the use of real-time updated digital twins for drone swarm command and control. We aim to enhance the effectiveness of drone swarm operations by introducing a novel method, Camera-based Adaptive Line Formation and Dynamic Leader-following Optimization (CALF-DLFO). We investigate the impact of this novel control method on area coverage and drone distribution, considering factors such as formation direction, speed, separation distance, and adjustments. Simulation experiments demonstrate a significant improvement in area coverage compared to prior methods which focused on isolated segments. Our camera-based adaptive line formation, combined with efficient drone control mechanisms, not only enhances coverage but also provides a promising approach to scalable and automated drone swarm management.

It is not uncommon for science fiction movies to portray futuristic user interfaces that can only be realized decades later with state-of-the-art technology. In this work, we present a prototypical augmented reality (AR) installation that was inspired by the movie The Time Machine (2002). It consists of a transparent screen that acts as a window through which users can see the stereoscopic projection of a three-dimensional virtual human (VH). However, there are some key differences between the vision of this technology and the way VHs on these displays are actually perceived. In particular, the additive light model of these displays causes darker VHs to appear more transparent, while light in the physical environment further increases transparency, which may affect the way VHs are perceived, to what degree they are trusted, and the distances one maintains from them in a spatial setting. In this paper, we present a user study in which we investigate how transparency in the scope of transparent AR screens affects the perception of a VH’s appearance, social presence with the VH, and the social space around users as defined by proxemics theory. Our results indicate that appearances are comparatively robust to transparency, while social presence improves in darker physical environments, and proxemic distances to the VH largely depend on one’s distance from the screen but are not noticeably affected by transparency. Overall, our results suggest that such transparent AR screens can be an effective technology for facilitating social interactions between users and VHs in a shared physical space.

Our paper presents LIFT, a system that enables educators to create immersive virtual field trip experiences for their students. LIFT overcomes the challenges of enabling non-technical educators to create their own content and allows educators to act as guides during the immersive experience. The system combines live-streamed 360° video, 3D models, and live instruction to create collaborative virtual field trips. To evaluate LIFT, we developed a field trip with biology educators from the University of Central Florida(UCF) and showcased it at a science festival. Our results suggest that LIFT can help educators create immersive educational content while out in the field. However, our pilot observational study at the museum highlighted the need for further research to explore the instructional design of mixed immersive content created with LIFT. Overall, our work provides an application development framework for educators to create immersive, hands-on field trip experiences.

Several critical limitations exist in the currently available tracking technologies for fully enclosed virtual reality (VR) systems. While several 6DOF tracking projects such as Hedgehog have successfully demonstrated excellent accuracy, precision, and robustness within moderate budgets, these projects still include elements of hardware that can interfere with the user's visual experience. The objective of this project is to design a tracking solution for fully enclosed VR displays that achieves comparable performance to available commercial solutions but without any artifacts that can obscure the user's view. JanusVF is a tracking solution involving a cooperation of both the hardware sensors and the software rendering system. A small, high-resolution camera is worn on the user's head, but faces backward (180 degree rotation about vertical from the user's perspective). After acquisition of the initial state, the VR rendering software draws specific fiducial markers with known size and absolute position inside the VR scene. These virtual markers are only drawn behind the user and in view of the camera. These fiducials are tracked by ARToolkitPlus and integrated by a single-constraint-at-a-time (SCAAT) filter algorithm to update the head pose. Experiments analyzing accuracy, precision, and latency in a six-sided CAVE-like system show performance that is comparable to alternative commercial technologies.

BACKGROUND: Linking high-throughput experimental data with biological networks is a key step for understanding complex biological systems. Currently, visualization tools for large metabolic networks often result in a dense web of connections that is difficult to interpret biologically. The MetNetGE application organizes and visualizes biological networks in a meaningful way to improve performance and biological interpretability. RESULTS: MetNetGE is an interactive visualization tool based on the Google Earth platform. MetNetGE features novel visualization techniques for pathway and ontology information display. Instead of simply showing hundreds of pathways in a complex graph, MetNetGE gives an overview of the network using the hierarchical pathway ontology using a novel layout, called the Enhanced Radial Space-Filling (ERSF) approach that allows the network to be summarized compactly. The non-tree edges in the pathway or gene ontology, which represent pathways or genes that belong to multiple categories, are linked using orbital connections in a third dimension. Biologists can easily identify highly activated pathways or gene ontology categories by mapping of summary experiment statistics such as coefficient of variation and overrepresentation values onto the visualization. After identifying such pathways, biologists can focus on the corresponding region to explore detailed pathway structure and experimental data in an aligned 3D tiered layout. In this paper, the use of MetNetGE is illustrated with pathway diagrams and data from E. coli and Arabidopsis. CONCLUSIONS: MetNetGE is a visualization tool that organizes biological networks according to a hierarchical ontology structure. The ERSF technique assigns attributes in 3D space, such as color, height, and transparency, to any ontological structure. For hierarchical data, the novel ERSF layout enables the user to identify pathways or categories that are differentially regulated in particular experiments. MetNetGE also displays complex biological pathway in an aligned 3D tiered layout for exploration.

Several critical limitations exist in the currently available tracking technologies for fully-enclosed virtual reality (VR) systems. While several 6DOF tracking projects such as Hedgehog [24] have successfully demonstrated excellent accuracy, precision, and robustness within moderate budgets, these projects still include elements of hardware that can interfere with the user's visual experience. The objective of this project is to design a tracking solution for fully-enclosed VR displays that achieves comparable performance to available commercial solutions but without any artifacts that can obscure the user's view. JanusVF is a tracking solution involving a cooperation of both the hardware sensors and the software rendering system. A small, high-resolution camera is worn on the user's head, but faces backwards (180 degree rotation about vertical from the user's perspective). After acquisition of the initial state, the VR rendering software draws specific fiducial markers with known size and absolute position inside the VR scene. These virtual markers are only drawn behind the user and in view of the camera. These fiducials are tracked by ARToolkitPlus [25] and integrated by a single-constraint-at-a-time (SCAAT) [26] filter algorithm to update the head pose. Early experiments in a six-sided CAVE-like system show performance that is comparable to alternative commercial technologies.

The goal of this project is to develop a training system that can simulate the welding process in real-time and give feedback that avoids learning wrong motion patterns for beginning welders and can be used to analyze the process by the teacher afterwards. The system is based mainly on COTS components. A standard PC with a Dual-core CPU and a medium-end nVidia graphics card is sufficient. Input is done with a regular welding gun to allow realistic training. The gun is tracked by an OptiTrack system with 3 FLEX:V100 cameras. The same is also used to track a regular welding helmet to get accurate eye positions for display, which was chosen over glasses for robustness. The display itself is a Zalman Trimon stereo monitor that is laid out horizontally. The software is designed around a main simulation component for solving heat conduction on a grid of simulation points based on local GaussSeidel elimination.

Multi-frame rate rendering is a parallel rendering technique that renders interactive parts of the scene on one graphics card while the rest of the scene is rendered asynchronously on a second graphics card. The resulting color and depth images of both render processes are composited and displayed. This paper presents advanced multi-frame rate rendering techniques, which remove limitations of the original approach and reduce artifacts. The interactive manipulation of light sources and their parameters affects the entire scene. Our multi-GPU deferred shading splits the rendering task into a rasterization and lighting pass and distributes the passes to the appropriate graphics card to enable light manipulations at high frame rates independent of the geometry complexity of the scene. We also developed a parallel volume rendering technique, which allows the manipulation of objects inside a translucent volume at high frame rates. Due to the asynchronous nature of multi-frame rate rendering artifacts may occur during the migration of objects from the slow to the fast graphics card, and vice versa. We show how proper state management can be used to avoid these artifacts almost completely. These techniques were developed in the context of a single-system multi-GPU setup, which considerably simplifies the implementation and increases performance.

We introduce a new concept for improved interaction with complex scenes: multi-frame rate rendering and display. Multi-frame rate rendering produces a multi-frame rate display by optically or digitally compositing the results of asynchronously running image generators. Interactive parts of a scene are rendered at the highest possible frame rates while the rest of the scene is rendered at regular frame rates. The composition of image components generated with different update rates may cause certain visual artifacts, which can be partially overcome with our rendering techniques. The results of a user study confirm that multi-frame rate rendering can significantly improve the interaction performance while slight visual artifacts are either not even recognized or gladly tolerated by users. Overall, digital composition shows the most promising results, since it introduces the least artifacts while requiring the transfer of frame buffer content between different image generators

Intrusion detection systems have been popular tools in the battle against adversaries who, for whatever reason, desire to break into networks, compromise hosts, and steal valuable information. One problem with current implementations, however, is the sheer number of alerts they can generate, many of which tend to be false alarms. This drawback makes effective use of such systems a challenging task. In this paper we explore three-dimensional approaches to visualizing network intrusion detection system alerts and aggregated network statistics in order to provide the system administrator with a better picture of the events occurring on his or her network. While some research has been done using two-dimensional concepts, 3D approaches have not received much attention with regard to detecting network intrusions. Evaluation of our visualizations using the 1999 DARPA intrusion detection evaluation data set demonstrates the potential benefit of utilizing the third dimension. We show how a number of attack types in the data set generate visual evidence of abnormal activity that a security administrator might use as motivation for further investigation.