Raymond H. Cuijpers

Fostering human–robot bonding is increasingly recognized as an important goal in human–robot interaction design, yet little is still known about the antecedents and consequences of such bonds. This study investigated caregiving toward robots—defined here as providing emotional support to a robot that appears distressed—as a potential antecedent of bonding, and examined whether bonding was related to the robot’s persuasive effectiveness. Individual characteristics were also considered as potential moderators. In a between-subjects experiment, participants played a game with a robot, during which they either provided care to it (n = 30) or did not (n = 29). After the interaction, the robot attempted to persuade participants to stand up and walk around the room. Bonding was assessed using a set of psychological scales, and behavioral compliance with the robot’s persuasion was video-recorded and coded. The results showed that caregiving did not, in itself, increase bonding. However, participants with higher trait empathy reported stronger bonding after providing care to the robot. We also found that caregiving significantly increased compliance with the robot’s persuasion, and that it appeared to create a relational context in which bonding and compliance became positively associated. These findings shed light on the mechanisms underlying social bonding and social influence in human–robot interaction.

Dynamic Neural Fields (DNFs) offer a biologically grounded framework for modelling continuous-time neural dynamics underlying a wide range of cognitive and sensorimotor functions. Despite their expressive power, DNF architectures are typically designed and tuned manually, as their nonlinear dynamics and kernel-based interactions make systematic design and parameterisation difficult. As a result, constructing functional DNF models requires expert knowledge and iterative trial-and-error. We introduce NEAT-DNFs, a neuroevolutionary framework that extends NeuroEvolution of Augmenting Topologies (NEAT) to the automated synthesis of DNF architectures. In NEAT-DNFs, neural fields and inter-field interactions are encoded as evolvable genes, enabling the joint evolution of intrinsic parameters and architectural topology. This allows continuous-time neural dynamics to be discovered autonomously, without manual tuning. We evaluate NEAT-DNFs on a hierarchy of benchmark tasks of increasing complexity. Initial tasks target core Dynamic Field Theory (DFT) mechanisms, including detection, working memory, and selection. We then consider more demanding problems that require the emergence of intermediate neural fields, thereby testing the framework's capacity for structural innovation. Across all tasks, NEAT-DNFs reliably evolves compact and stable architectures, increasing structural complexity only when necessary. This work bridges neuroevolution and DFT, establishing a foundation for the automated design of biologically inspired neural field models.

Achieving fluent and adaptive Human–Robot Interaction (HRI) demands control architectures that generate behaviour through continuous sensorimotor dynamics. Dynamic Neural Fields (DNFs) offer a biologically inspired framework for modelling perception, memory, and action selection, yet designing functional neural-field controllers remains largely manual, limiting their scalability and broader adoption in adaptive robotic systems.Here, we introduce an approach for the automatic synthesis of DNF controllers capable of generating joint-action behaviours. Extending the NeuroEvolution of Augmenting Topologies (NEAT) algorithm to operate on neural-field-based controllers, the approach evolves both intrinsic field dynamics and architectural topology through mutation, crossover, and speciation. We evaluate it in a collaborative packaging task designed to isolate three fundamental joint-action behaviours (complementary action selection, cooperative assistance, and action inhibition) and whose minimal controller can be derived analytically, providing a ground-truth reference for assessing evolved solutions. Across 100 independent runs, the approach discovered viable controllers in 97 cases. The evolved architectures were topologically minimal, with interpretable activation dynamics grounded in known Dynamic Field Theory (DFT) mechanisms. The highest-performing minimal controller was deployed directly to a physical robot without parameter retuning, generalising to object arrangements beyond those used during evolution. Evolution converged on the minimal architecture without any specification of intermediate structure, indicating that behavioural requirements can drive the emergence of the internal processing structure the task requires. These findings demonstrate that functional DNF architectures for joint action can be evolved rather than hand-designed, providing a constructive pathway toward bio-inspired, automatically synthesised control for adaptive HRI.

Social robots are increasingly applied as health behavior change interventions, yet actionable knowledge to guide their design and evaluation remains limited. This systematic review synthesizes (1) the behavior change strategies used in existing HRI studies employing social robots to promote health behavior change, and (2) the evaluation methods applied to assess behavior change outcomes. Relevant literature was identified through systematic database searches and hand searches. Analysis of 39 studies revealed four overarching categories of behavior change strategies: coaching strategies, counseling strategies, social influence strategies, and persuasion-enhancing strategies. These strategies highlight the unique affordances of social robots as behavior change interventions and offer valuable design heuristics. The review also identified key characteristics of current evaluation practices, including study designs, settings, durations, and outcome measures, on the basis of which we propose several directions for future HRI research.

The pupil response serves as a proxy for the biological potency of light. A reproducible measurement of the pupil response requires controlled dark adaptation, but it is currently unclear whether shorter dark adaptation periods can be used to reliably determine the biological potency of light through measuring the pupil response. Therefore, we conduct a within-participant experiment, continuously measuring the pupil response for dark adaptation periods between 40 and 120 seconds. Results show a trend in the effect of mEDI on the 6s-PIPR increasing with longer adaptation periods. The dark adaptation period of 120 seconds does not fit this trend, and more research is needed here. Overall, our study shows that a dark adaptation period of 60 to 100 seconds is suitable for repeated measurements of the 6s-PIPR, with the aim of using the pupil response as a proxy for the biological potency of light.

By providing daytime structure, Socially Assistive Robots (SARs) can assist older adults in autonomy and mitigate global challenges associated with the ageing society. Despite this potential, SARs are not yet fully integrated into healthcare. Therefore, this research explores the effectiveness of personalisation in communication and language use of SARs, using mass customization and personalised persuasion, to enhance acceptance and perceived persuasiveness of SARs for older adults. The research comprised two studies. Study 1 examined needs in communication and language use of SARs among older adults (N = 6), while Study 2 (N = 33) investigated the impact of mass customization on acceptance and perceived persuasiveness, along with personalised persuasion on perceived persuasiveness and suggestion adoption in a laboratory setting. Study 1 identified six customizable variables, including speech characteristics and the content of messages. Study 2 showed a significant positive effect of mass customization on perceived persuasiveness. Additionally, positive but non-significant correlations were observed between mass customization and acceptance, as well as between personalised persuasion and perceived persuasiveness. Therefore, this study suggests that personalisation through mass customization in SAR communication positively impacts the effectiveness of interventions for older adults. Accordingly, integrating customization in SAR communication may enhance interaction with older adults, addressing healthcare challenges associated with the ageing society.