Self-monitoring, a commonly used technique in cognitive-behavioral interventions, involves the objective observation and organization of one’s internal states. However, providing continuous, individualized support for this process is challenging in resource-limited psychiatric daycare settings. While social robots show promise in facilitating self-disclosure, evidence regarding their long-term practical utility and clinical relevance in real-world clinical environments remains insufficient. To address this gap, we conducted a two-phase field study comprising periodic sessions held over a 24-month period to develop and evaluate a social robot-assisted self-monitoring support system. In the 14-month Development Phase (N = 39), we employed a practice-informed design process to establish robot-specific design principles, prioritizing a non-judgmental listener role over advice-giving and offering user-selected positive and negative topics to accommodate diverse psychosocial needs. This was followed by a 10-month Evaluation Phase (N = 20), consisting of 10 experimental days, to assess the finalized system’s impact. Key findings indicated that: (1) single dialogue sessions were followed by significant improvements in participants’ subjective overall health and alertness; (2) participants demonstrated a strong preference for positive-oriented topics; and (3) qualitative analysis revealed emergent reflective insights and perceived positive changes. These results provide empirical evidence suggesting the system’s practical utility and clinical relevance, highlighting the value of a user-centered approach that empowers patients to choose topics aligned with their recovery goals.
This paper presents the design and evaluation of a robotic rollator, FALL-E, developed to investigate how individuals respond to perturbations during overground walking with an assistive device. Drawing inspiration from perturbation methods used in treadmill and kinesthetic haptic studies, we introduce a novel overground perturbation technique that delivers brief, directional changes in velocity through the motorized rear wheels of FALL-E. We characterize the resulting stabilizing responses in able-bodied individuals at low risk of falls. Our findings show that changes in step length and anterior-posterior margin of stability (AP MOS) consistently aligned with the direction of the perturbation. In addition, assisted walking with FALL-E elicited distinct biomechanical adaptations not seen in unassisted conditions, including modifications in elbow flexion and trunk flexion. These results provide new insight into balance control strategies during perturbed, device-assisted walking.
Public-facing social robots, including robotic comedians, need to recognize onlooker reactions and repair failures when an interaction falls flat. In the presented work, we use stand-up comedy as a structured and fast-paced testbed to partially address this broader challenge of perceiving and adapting to failures; the robot can deliver a silly joke, observe an immediate audience response, and select an appropriate follow-up within a short interaction loop. We compare three room-reading strategies for informing post-joke adaptation by a fully autonomous NAO robot comedian. The strategies use crowd audio cues, visual facial reactions, or spoken audience feedback to select repair-oriented responses. We evaluated these three condition approaches in 19 total live performances (split between conditions) with \(N\) =80 total participants and measured experience using performance-level joke ratings, perceived social closeness, and social attribute ratings, along with post-performance group interviews. Across conditions, outcomes were broadly positive and we found no statistically significant differences between strategies, which suggests that each modality can support viable adaptation while capturing different forms of audience engagement. The study's manipulation check and interviews also showed that presumed robot adaptivity mode mattered, especially for the verbal strategy, which could be compelling when audiences engaged but was less reliably recognized when groups were quiet. This work supports flexible room-reading tactics for failure repair in public-facing robots while highlighting the importance of making a robot's adaptive behavior understandable in the moment.
The increasing deployment of Automated Mobile Vehicles (AMVs) in warehouses requires efficient navigation in dynamic environments shared with human workers. Traditional safety strategies, such as stopping or slowing down, often reduce efficiency and do not account for human-robot interaction. To address this, interaction-aware methods have been developed. However, traditional interaction models assume fixed parameters, rely on a single forecast, or use intent labels that are not updated during the interaction, and thus do not explicitly represent intention uncertainty, that is, variability in how people react to robots. When combined with position uncertainty introduced by sensing noise, this unmodeled intention uncertainty can make motion prediction and navigation decisions unreliable, leading to conservative behavior, inefficiency, or collisions in human–robot interactions. To bridge this gap, we propose Belief-based Risk-Aware Anticipatory Navigation (BRAN), a framework that jointly models intention and position uncertainty. BRAN maintains probabilistic beliefs over an agent's position and intention to explicitly represent uncertainty in both quantities. Position uncertainty is represented via sampled observations, and the intent belief is updated online through Bayesian inference so that predictions evolve with the interaction. These beliefs are propagated through the Social Force Model (SFM) to generate predictions of possible agent motion under uncertainty. We then compute collision risk from the predicted motion distribution and incorporate it into model predictive control as a constraint. This constrains predicted collision risk while optimizing motion toward the goal. Experimental results validate the proposed design in controlled settings, demonstrate its effectiveness in scenarios with multiple pedestrians, and confirm its feasibility on a real robot with onboard sensing and actuation.
Human-robot interaction (HRI) failures remain a major barrier to deploying robots in real-world environments. Prior work often treats failures as isolated technical faults or focuses on post-hoc recovery behaviors. In practice, many breakdowns arise because humans and robots operate under inconsistent assumptions about the current world state. We propose WSM-Aware HRI, an IoT-enhanced modular framework that unifies diverse HRI breakdowns as World-State Mismatches (WSMs) between a human's instruction-implied assumptions and a robot's grounded world model built from multimodal perception and digital augmentation. A Large Language Model (LLM) is used to make implicit assumptions explicit, map them to a small set of mismatch types, and specify the evidence needed for verification against the robot's world state. WSM-Aware HRI shifts failure handling from execution-time recovery to proactive mismatch detection during intention formation, enabling interventions guided by safety, norm compliance, and multi-user coordination with transparent explanations. We evaluate mismatch identification in ten everyday cases spanning both visual and latent-state mismatches. The system can accurately produce the expected output results, and ablations show that reliable identification depends on appropriate grounding representations and verification-oriented refinement. These results indicate that treating interaction breakdowns as explicit world-state mismatches enables earlier detection of impending failures and offers a principled mechanism for integrating external evidence and social constraints into human-robot interaction.
