This paper explores active sensing strategies that employ vision-based tactile sensors for robotic perception and classification of fabric textures. We formalize the active sampling problem in the context of tactile fabric recognition and provide an implementation of information-theoretic exploration strategies based on minimizing predictive entropy and variance of probabilistic models. Through ablation studies and human experiments, we investigate which components are crucial for quick and reliable texture recognition. Along with the active sampling strategies, we evaluate neural network architectures, representations of uncertainty, influence of data augmentation, and dataset variability. By evaluating our method on a previously published Active Clothing Perception Dataset and on a real robotic system, we establish that the choice of the active exploration strategy has only a minor influence on the recognition accuracy, whereas data augmentation and dropout rate play a significantly larger role. In a comparison study, while humans achieve 66.9% recognition accuracy, our best approach reaches 90.0% in under 5 touches, highlighting that vision-based tactile sensors are highly effective for fabric texture recognition.

In robotics, understanding human interaction with autonomous systems is crucial for enhancing collaborative technologies. We focus on human-swarm interaction (HSI), exploring how active robot groups of varying sizes affect operators’ cognitive and perceptual reactions over distinct durations. We analyze the impact of different numbers of active robots within a 15-robot swarm on operators’ time perception, emotional state, flow experience, and task difficulty perception. Our findings indicate that managing multiple active robots when compared to one active robot significantly alters time perception and flow experience, leading to a faster passage of time and increased flow. More active robots and extended durations cause increased emotional arousal and perceived task difficulty, highlighting the interaction between the number of active robots and human cognitive processes. These insights inform the creation of intuitive human-swarm interfaces and aid in developing swarm robotic systems aligned with human cognitive structures, enhancing human-robot collaboration.

Many large-scale multi-robot systems require human input during operation in different applications. To still minimize the human effort, interaction is intermittent or restricted to a subset of robots. Despite this reduced demand for human interaction, the mental load and stress can be challenging for the human operator. A specific effect of human-swarm interaction may be a hypothesized change of subjective time perception in the human operator. In a series of simple human-swarm interaction experiments with robot swarms of up to 15 physical robots, we study whether human operators have altered time perception due to the number of controlled robots or robot speeds. Using data gathered by questionnaires, we found that increased swarm size shrinks perceived time and decreased robot speeds expand the perceived time. We introduce the concept of subjective time perception to human-swarm interaction. Future research will enable swarm systems to autonomously modulate subjective timing to ease the job of human operators.

When small holes are felt with the tongue they are perceived to be larger as compared to when felt with the index finger. It is an open question why the size perceived with the tongue differs from size perceived with the finger. We suggest that differences in perceived size are due to differences in the effector's deformation at the edge of the explored hole, which correlate with the effector's pliability. The more pliable the effector is, the more it will deform and the hole will be perceived to be larger. In two experiments we tested this hypothesis. In Experiment I we manipulated the force applied by the tongue, and thus the tongue's deformation. As predicted, the tongue perceived holes to be larger when higher forces were applied. Experiment II investigated how the toe perceives size as compared to the tongue and finger. The toe is less pliable than the finger and the finger is less pliable than the tongue. Also as predicted, holes at the toe were perceived to be smaller than holes at the finger and considerably smaller than holes at the tongue. The findings corroborate our pliability-deformation hypothesis of haptic size perception.

An object's softness is stereotypically judged by pressure movements indenting the surface [1]. In exploration without movement constraints, participants repeat such indentation movements. We investigated how people modulate executed peak forces for different indentations depending on stimulus softness. Participants performed a 2AFC discrimination task for stimulus pairs from one of 4 softness categories. We assumed that movement control at different exploration moments is based on variations in the predictive and sensory signals available. We manipulated availability of predictive signals on softness category, by presenting either stimuli of the same category in a blocked fashion (high predictability) or by randomly mixing stimuli from different categories (low predictability). Effects of sensory signals were examined by contrasting first and last indentation, as sensory signals are hardly available when initiating exploration but gathered during exploration. The results show that participants systematically apply lower forces when sensory or predictive signals indicate softer objects as compared to harder objects. We conclude that softness exploration can be considered as a sensorimotor control loop, in which predictive and sensory signals determine movement control. Further, the results indicate a high importance of predictive processes throughout the entire exploration, as effects of predictive signals maintain in the last indentation.

The perception of softness is the result of the integration of information provided by multiple cutaneous and kinesthetic signals. The relative contributions of these signals to the combined percept of softness was not yet addressed directly. We transmitted subtle external vertical forces to the exploring human finger during the exploration of deformable silicone rubber stimuli to dissociate the force estimates provided by the kinesthetic signals and the efference copy from cutaneous force estimates. This manipulation introduced a conflict between the cutaneous and the kinesthetic/efference copy information on softness. We measured Points of Subjective Equality (PSE) of manipulated references to stimuli which were explored without external forces. PSEs shifted as a linear function of external force in predicted directions - to higher compliances with pushing and to lower compliances with pulling force. We found relative contribution of kinesthetic/efference copy information to perceived softness being 23% for rather hard and 29% for rather soft stimuli. Our results suggest that an integration of the kinesthetic/efference copy information and cutaneous information with constant weights underlies softness perception. The kinesthetic/efference copy information seems to be slightly more important for the perception of rather soft stimuli.

People perceive a smaller and denser object to be heavier than a larger, less dense object of the same mass. We developed a new model of heaviness perception that can explain this size-weight illusion. Modeling followed recent insights on principles of information integration. Perceived heaviness is modeled as a weighted average of one heaviness estimate derived from object mass and another one derived from object density with weights that follow estimate reliabilities. In an experiment, participants judged the heaviness of 18 objects using magnitude estimation methods. Objects varied in mass and density. We also varied the reliability of density information by varying visual reliability: Participants were blindfolded or had strongly impaired, mildly impaired or full vision. Because participants lifted each object via a string they required visual information on object size to assess object density. The pattern of mass and density influences on judged heaviness confirmed model predictions. Also as predicted, density influences on judged heaviness increased with increasing reliability, whereas mass influences decreased. Individual and average data were well fit by the model (r2s > 0.96). Density information contributed for 14%, 21 % and 29% to heaviness, when vision was strongly impaired, mildly impaired or not impaired, respectively. Overall, the results highly corroborate our model, which appears to be promising as unifying framework for a number of findings on the size-weight illusion.

The active control of exploratory movements is an integral part of active touch. We investigated how the spatio-temporal extension of an exploration affects texture discrimination via active touch and how participants spontaneously control the extension of their exploration. Participants stroke one to 8 times across a virtual grating. Half of the participants judged the grating according to ridge amplitude, the other half judged ridge period. In both tasks just noticeable differences (JNDs) decreased with an increasing number of strokes, down to a (presumably) minimum JND for stroke numbers of about 6-7 or higher (according to a model fit). This extends previous findings on the perceptual effects of extended stimulation to the domain of active touch. Further, for spontaneous exploration, 6-7 strokes may be regarded as being optimal, given that this extension results in minimal JNDs while using the smallest number of strokes necessary. Many participants tended to spontaneously use clearly more than this optimal number of strokes, before they had experience with the task. With experience, stroke numbers were, however, individually closer to and, on average slightly below this number. That is, participants seem to have adapted their exploration to the task. This supports the view that the active control of exploratory movements can be used in order to optimize performance-when participants have sufficient experience with the haptic task.

Based on existing knowledge on human tactile movement perception, we constructed a prototype of a novel tactile multipin display that controls lateral pin displacement and, thus produces shear force. Two experiments focus on the question of whether the prototype display generates tactile stimulation that is appropriate for the sensitivity of human tactile perception. In particular, Experiment I studied human resolution for distinguishing between different directions of pin displacement and Experiment II explored the perceptual integration of information resulting from the displacement of multiple pins. Both experiments demonstrated that humans can discriminate between directions of the displacements, and also that the technically realized resolution of the display exceeds the perceptual resolution (>14°). Experiment II demonstrated that the human brain does not process stimulation from the different pins of the display independent of one another at least concerning direction. The acquired psychophysical knowledge based on this new technology will in return be used to improve the design of the display.