The growing interest in remote collaboration is driven by multiple factors. Advances in digital technology have made it increasingly feasible to collaborate from anywhere, breaking down geographical barriers. Despite the rise of new tools several challenges still exist. These include limitations on the remote member's perspective, confined to what on-site members can capture, as well as potential shortcomings for interacting with on-site physical objects. This position paper discusses the potential of leveraging mobile robots within eXtended Reality (XR) environments to transform how remote collaboration is performed. Mobile robots, equipped with cameras, microphones, articulated arms and various sensors, can serve as physical avatars for remote team members, allowing to navigate through physical spaces, enabling real-time, immersive experiences. Combined with XR tools, these robots may empower on-site and remote collaborators to engage in shared activities. These and other arguments in favor of this position are presented and future directions are proposed.

In human communication we need to establish the target of our message and understand if someone is addressing us. Such mechanisms facilitate communication in environments where several potential interlocutors exist. With the advances of speech technologies supporting interaction with computers, the establishment of a device as an interlocutor has often been performed resorting to wake-up words. While this addresses the issue, it is far from the naturalness and efficiency of what we can accomplish in human-human communication. In this regard, research has considered alternatives, such as the visual focus of attention, but the implementations are often scenario specific and not easily available for a generalized use. In this paper, we argue that the establishment of a machine as an interlocutor, particularly for speech interaction, should consider a wide range of verbal and nonverbal aspects, and we conceptualize and present a first proof-of-concept of its integration as a core feature in a multimodal interactive framework.

Currently, spoken dialogue systems mostly consider only voice interactions and do not consider nonverbal cues. This work aims to explore the domain of interaction with conversational assistants, taking into account not only verbal cues, but also the importance of integrating nonverbal aspects to make interactions between humans and agents more natural, richer, and more efficient. This work presents an initial proof-of-concept for the modules needed for dialogue systems to take into account verbal and nonverbal inputs, as well as their articulation, demonstrating their viability. To this end, we use a fusion module that can directly or indirectly affect the communicative acts, depending on the verbal and nonverbal communication inputs. Preliminary results demonstrate the implemented system behaviour of how verbal content is affected by nonverbal cues.

For aging speech, there is limited knowledge regarding the articulatory adjustments underlying the acoustic findings observed in previous studies. In order to investigate the age-related articulatory differences in European Portuguese (EP) vowels, the present study analyzes the tongue configuration of the nine EP oral vowels (isolated context and pseudoword context) produced by 10 female speakers of two different age groups (young and old). From the tongue contours automatically segmented from the US images and manually revised, the parameters (tongue height and tongue advancement) were extracted. The results suggest that the tongue tends to be higher and more advanced for the older females compared to the younger ones for almost all vowels. Thus, the vowel articulatory space tends to be higher, advanced, and bigger with age. For older females, unlike younger females that presented a sharp reduction in the articulatory vowel space in disyllabic sequences, the vowel space tends to be more advanced for isolated vowels compared with vowels produced in disyllabic sequences. This study extends our pilot research by reporting articulatory data from more speakers based on an improved automatic method of tongue contours tracing, and it performs an inter-speaker comparison through the application of a novel normalization procedure.

The evolution of human-machine interfaces has resulted in the proposal of increasingly complex multimodal solutions, encompassing a diverse set of input and output modalities, towards more natural and efficient interaction paradigms. Supporting the required range of interaction possibilities entails the development of complex systems which, as the number of features and involved technologies grows, pose challenges to research and development. This arises from the inherent difficulty in understanding how the system is working and how specific modules are behaving, important during development, integration, and testing, but also during deployment of these systems for evaluation by users, whether in a lab or in real environments. For these scenarios, approaches based on conventional logging or post-task questionnaires are no longer viable. In response, a first stage to tackle these challenges is proposed based on a systematic approach to monitoring interaction in these multimodal architectures. To this end, this work presents and illustrates the use of MNEMOSINE, the first instantiation of a framework offering real-time monitoring capabilities and event storage mechanisms for multimodal architectures. MNEMOSINE can serve as grounds for developers and researchers to get a structured view over current system status and creates the conditions supporting the analysis of system behavior over time. This aims to enhance the understanding and supervision of interactions within multimodal systems and can, in turn, foster improved system development and refinement, subsequently supporting the integration of increasingly multimodal systems into daily life.