Touching a Cat Without Touch: Does Mid-Air Ultrasound Haptic Feedback Promote Relaxation in Virtual Cat Interaction?
Juro Hosoi, Yuki Ban, Shin'ichi Warisawa · 2026 · Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems (CHI '26) · doi:10.1145/3772318.3790818
Summary
Hosoi and colleagues investigate whether mid-air ultrasound haptics — tactile sensations produced by focused acoustic radiation pressure, with no device worn on the user's skin — can convey the feel of a virtual cat convincingly enough to support relaxation in VR. The paper comprises two studies. A perceptual study (n=21) uses a 2x2 within-subjects design to ask how periodic modulation of the stimulated area and the ultrasound intensity (via circular spatio-temporal modulation over the palm) communicates a cat's breathing. Participants rate conditions on the Haptic Experience (HX) model's five dimensions (autotelics, realism, harmony, expressivity, immersion). The authors then build a VR relaxation application in which a seated participant, wearing a Meta Quest 3 and noise-cancelling headphones, interacts with a virtual cat that breathes, purrs, yawns, and responds to the tracked hand. A second application study (n=22/21) compares three conditions after a 2-back cognitive-load task: no haptics (None), Meta Quest Touch Plus controller vibration (Controller), and mid-air ultrasound haptics delivered by two Ultraleap AUTDs combined with Leap Motion hand tracking (Ultrasound). Dependent measures are heart rate and ln(RMSSD) from ECG, Self-Assessment Manikin valence and arousal, and custom Likert items for realism, attachment, empathy, agency, and enjoyment. The paper positions itself at the intersection of HCI, haptics, and human-animal interaction, arguing that the contact-free nature of ultrasound can deliver biologically plausible rhythmic cues (breathing, fur stroking) without the hygiene, setup, or wearable burden that has limited prior haptic creatures and animal-inspired robots.
Key findings
In the perceptual study, periodic area modulation was the dominant cue for conveying a cat's breathing: area had significant main effects on all five HX measures, while amplitude produced significant effects only on realism and expressivity, with an Amplitude x Area interaction on expressivity (p=0.034). In the application study, only the Ultrasound condition yielded a significant pre/post reduction in heart rate (p=0.010, Cohen's d=0.72, approximately -2 bpm / -3%), indicating suppression of sympathetic arousal after the cognitive-load task. ln(RMSSD) did not change significantly in any condition, consistent with the three-minute recovery being too short to produce a full parasympathetic rebound. On subjective measures, Ultrasound scored significantly higher than None on all five items (realism, attachment, empathy, agency, enjoyment), with medium-to-large effects, and showed medium-sized advantages over Controller on attachment, empathy, and agency. In a forced-choice preference question, 72.7% of participants preferred the Ultrasound condition; they cited the freedom of finger movement without holding a controller, the naturalness of the gentle tactile sensation, and the novelty of the experience.
Relevance
Although the paper frames itself around VR relaxation and animal-assisted therapy, the more interesting accessibility angle is that mid-air ultrasound haptics renders spatially resolved tactile information in the air, with no worn device, no held controller, and no contact surface. That combination is promising for blind and low-vision users, where current haptic options force a trade-off between resolution (refreshable braille, wearables) and freedom of movement (controller vibration). The finding that hand-tracked ultrasound produced higher agency ratings than a controller is directly relevant: controllers occupy the hands and occlude fine finger gestures that BLV users rely on for exploration. Limitations include the short session duration, the fixed seated posture, and the expense and calibration burden of current ultrasound arrays (~USD 4,000 for research kits). Future work could evaluate ultrasound as a non-visual output channel for tactile graphics, wayfinding cues, or document navigation.
Tags: mid-air haptics · ultrasound haptics · virtual reality · haptic feedback · human-animal interaction · relaxation · affective computing · non-contact interaction · multimodal interaction · sensory substitution