Tactile Search: Enhancing Targeting in 3D Space

1 INTRODUCTION

• Haptics are motivated as a complementary spatial channel when sensory resources are overloaded, divided, or degraded, conditions associated with declining search performance.Prior work also reports overload-related effects in navigation and driving.Source: §1 · p. 2
• Existing haptic spatial guidance can shift location information away from visual load, use available controllers without extra hardware, and support accessibility and safety contexts.The passage notes possible relevance for deaf and hard-of-hearing users, blind users, and safety-critical XR applications.Source: §1 · p. 2
• The study extends prior amplitude-based static horizontal localization to dynamic visual search and adds vertical position through frequency modulation.Dynamic search is treated as distinct from static localization.Source: §1 · p. 2
• The stated research questions cover effects on performance and confidence, vertical-position localization, and search behaviors including turning, switching, and overshooting.The study also frames its task as a controlled abstraction of real-world dynamic search and guidance demands.Source: §1 · p. 2
• The approach uses handheld-controller amplitude and frequency modulation in a competitive visual-search game, evaluated through public demonstration and controlled laboratory analysis.The work examines performance and behavioral responses to spatial haptic signals.Source: §1 · p. 2
• The contributions report faster targeting and higher confidence across a public study and lab experiment, more even performance across heights with combined haptics, and changes in search behavior.The behavioral analysis reports fewer unnecessary switches and overshoots and improved initial turn direction.Source: §1 · p. 2

2 RELATED WORK

• The related work is organized around sensory load and multimodal search, haptic guidance when other modalities are busy or unreliable, and sensory substitution.These areas motivate alternative channels for spatial information.Source: §2 · p. 2

2.1 Sensory Overload and Multimodal Search

• When concurrent tasks compete for shared resources, spatial-task performance can decline; distributing selected information across modalities may improve efficiency.Examples include smart-glasses walking tasks and proposed tactile in-vehicle displays.Source: §2.1 · pp. 2–3

2.2 Haptic Feedback for Spatial Guidance and Search

• Prior haptic guidance has supported navigation, spatial awareness, and search by encoding direction, distance, or elevation, though many systems use specialized hardware or constrained tasks.Examples include ultrasound and wearable or contactless guidance systems.Source: §2.2 · p. 3

2.3 Sensory Substitution and Crossmodal Spatial Perception

• Sensory-substitution systems show that spatial structure can be remapped across modalities, and prior handheld-controller amplitude differences enabled static horizontal localization in VR.Such mappings can be learned when they are consistent and preserve spatial structure.Source: §2.3 · p. 3

3 TACTILE SEARCH

• Tactile Search extends the handheld amplitude method in with frequency modulation for vertical position, using two controllers to localize objects across 3D space.The approach is evaluated in a competitive VR search game under naturalistic search conditions.Source: §3 · p. 3

3.1 Design and Implementation

• Amplitude varies with the controllers’ relative distance to the target, emphasizing differences at closer ranges through a nonlinear scaling curve.The method simulates interaural level differences used in auditory localization.Source: §3.1 · p. 3
• Higher frequencies encode elevated targets and lower frequencies encode lower ones; the implementation uses 125 Hz at eye level, 90 Hz low, and 150 Hz high.The three-height implementation is optimized for targets at a fixed distance of three meters.Source: §3.1 · p. 3

3.2 Game Implementation

• The game teaches vibration and frequency-height mappings, then asks users to find a blue sphere among 80 circling distractors and select it with a crosshair and controller button.Targets appear behind the user at low, middle, or high positions and move clockwise or counterclockwise.Source: §3.2 · pp. 3–4, Figure 2 · p. 4
• After two warm-up trials, each session presents 30 randomized trials, balanced across haptics and no-haptics conditions and the three target heights.Users receive a post-session comparison of their performance across the two conditions.Source: §3.2 · p. 4

4.1 Participants

• The conference demonstration enrolled 55 consenting passersby, but its public setting did not permit collection of demographic information.The demonstration aimed to assess whether spatialized vibrotactile signals improved visual-search performance.Source: §4.1 · p. 4, §4 · p. 4

4.2 Ethics; 4.3 Experimental Design

• All participants gave informed consent under institutionally approved procedures, and the within-subjects design varied haptic feedback and target height.Duration was the dependent performance measure.Source: §4.2 · p. 4, §4.3 · p. 4

4.4 Procedure

• The demo introduced sensory-substitution localization in VR, used a public scoreboard to incentivize participation, and treated controller audio as masked by the noisy setting.Participants followed the game procedure after a brief introduction and headset adjustment.Source: §4.4 · p. 4

4.5 Measurements

• The public study measured seconds to find and select a sphere and confidence on a 10-point agreement scale.Confidence ratings were collected for conditions with and without haptics.Source: §4.5 · p. 4

4.6.1 Confidence Ratings

• Confidence averaged 8.58 with vibrations versus 6.40 without, and participants were significantly more confident when vibrotactile cues were active (Figure 3c).The paired comparison was significant (Z = −4.968, p < .001).Source: §4.6.1 · pp. 4–5, Figure 3 · p. 5

4.6.2 Performance Ratings

• 3.62 seconds (SD = 1.70) with haptics versus 4.23 seconds (SD = 2.39) without; haptic trials were significantly faster (Figure 3a).The haptics main effect was F(1,54) = 34.47, p < .001, η2p = .390.Source: §4.6.2 · p. 5, Figure 3 · p. 5
• Medium-height targets were faster (M = 3.70) than high (M = 4.07) or low targets (M = 4.02); high and low did not differ (Figure 3b).The height effect was quadratic, with medium-height search faster than both other heights.Source: §4.6.2 · p. 5, Figure 3 · p. 5

5 EXPERIMENT II: LAB STUDY

• The lab study tested whether spatialized vibrotactile cues improve vertical visual search performance and behavior (RQ2 and RQ3).The study was conducted in a controlled environment.Source: §5 · p. 5

5.1 Participants

• Twenty-eight participants took part; a power analysis estimated approximately .98 power to detect a haptics effect comparable to Study 1 at α = .05.Participants were aged 21–39, and the estimated effect size from Study 1 was approximately 0.8.Source: §5.1 · pp. 5–6

5.2 Ethics

• The study followed institutional ethical guidelines, received ethics committee approval, and obtained written informed consent from participants.Source: §5.2 · p. 6

5.3 Experimental Design

• All participants completed a within-subjects competitive visual search task varying haptic cues and target height, with target-selection time as the dependent measure.The two factors were haptics versus no haptics and low, medium, or high target height.Source: §5.3 · p. 6

5.4 Procedure

• Participants onboarded, then searched for vibrating and non-vibrating targets as quickly as possible; afterward, they rated confidence and haptics and gave qualitative feedback.The researchers noted a slight concern about faint controller buzzes, which they considered too subtle to provide a reliable cue.Source: §5.4 · p. 6

5.5 Measurements

• Measurements included trial duration, head and controller movement, in-game events and sphere locations, confidence, perceived haptics contribution, and qualitative feedback.User and trial tracking were logged at 8 Hz; confidence used a 10-point scale and perceived haptics contribution a 7-point scale.Source: §5.5 · p. 6

• Haptic feedback significantly improved performance and increased participants’ subjective confidence in the lab study.Source: §6 · p. 6
• With combined haptic cues, performance did not differ across target heights, unlike the height-related performance difference observed without haptics.Source: §6 · p. 6
• Haptics reduced directional switches and overshoots and were associated with a higher probability of correct initial turns toward the target.Source: §6 · p. 6
• The authors also report qualitative feedback about participants’ impressions of the cues.Source: §6 · p. 6

• 8.18 average confidence with vibrations, compared with 5.25 without vibrations.The respective medians were 8 and 5.Source: §6.1 · p. 6
• The confidence difference between haptics and no-haptics conditions was significant (Z = −4.343, p < .001).Source: §6.1 · p. 6
• The lab confidence result reinforced the finding from the public demonstrator that participants were more confident with vibrotactile cues.Source: §6.1 · p. 6

6.2.1 Main Effects

• 3.79 seconds with haptics versus 4.11 seconds without haptics; the haptics main effect was significant (F(1,27) = 4.94, p = .035, η2p = .155).Reported durations are based on Box–Cox-transformed data; the analysis used a two-way repeated-measures ANOVA.Source: §6.2.1 · p. 6, §6.2 · p. 6
• Overall mean durations were 3.89 seconds at medium height, 3.91 seconds at high height, and 4.05 seconds at low height; the height main effect was not significant (p = .056).Source: §6.2.1 · p. 6

6.2.2 Interaction Effects

• The haptics-by-height interaction was significant (F(2,26) = 3.48, p = .046); without haptics, low targets were slower than medium and high targets.At low height, haptics improved performance (p < .001), while no haptics difference was found at medium or high heights.Source: §6.2.2 · pp. 6–7
• With haptics, performance did not differ among height levels (all p = 1.000), so no low-height performance disadvantage was observed in that condition.The reported p values are Bonferroni-adjusted pairwise comparisons.Source: §6.2.2 · p. 7

6.3 Search Behavior

• Search behavior was quantified from logged events and positions, with Box–Cox transformations and corrected follow-up comparisons.The analyzed measures included horizontal and vertical directional switches, target overshoots, and initial turns.Source: §6.3 · p. 7

6.3.1 Horizontal Switches

• Horizontal switches showed no overall haptic effect, but a haptics-by-height interaction: without haptics, low targets elicited more switches than medium or high targets, while haptics removed these differences.Height had a main effect, but corrected pairwise comparisons did not find significant differences between heights.Source: §6.3.1 · p. 7

6.3.2 Vertical Directional Changes

• Vertical switches varied by height, with low and high targets differing, but showed neither an overall haptic effect nor a haptics-by-height interaction.The low-versus-medium and medium-versus-high comparisons were not significant.Source: §6.3.2 · p. 7

6.3.3 Target Overshoots

• Overshoots showed no overall haptic effect, but a haptics-by-height interaction: haptics reduced low-target overshoots, while without haptics low targets exceeded medium and high targets.Overshoots also varied by height overall, with more for high than medium targets in pairwise comparisons.Source: §6.3.3 · p. 7

6.3.4 Correct Turns

• Haptics improved correct first turns by about 8% on average, with a 99.7% posterior probability of benefit.The results provide strong evidence that haptics helps participants initiate turns in the correct direction.Source: §6.3.4 · p. 8

6.4 Qualitative Analysis

• Horizontal localization was described as intuitive and reliable in 35 mentions of participant responses.Responses characterized vibrations as helping users orient left and right and find a general direction.Source: §6.4 · p. 8
• Vertical guidance drew 34 mentions spanning four patterns, including informative cues, difficulty judging height, and continued reliance on vision.Some participants found height cues unintuitive or difficult to distinguish.Source: §6.4 · p. 8
• A small group said vertical feedback required more familiarization, training, or exposure time.Some participants reported that the height cues took getting used to.Source: §6.4 · p. 8

6.5 Summary

• Haptic feedback significantly reduced task duration and increased users’ confidence in localizing targets.The lab study found an overall task-duration benefit compared with trials without haptics.Source: §6.5 · p. 8
• Without haptics, low-height performance was worse than at medium and high heights; in the combined haptic condition, all heights performed comparably.Similar height-related patterns appeared for horizontal switches and target overshoots, while vertical switches varied with height.Source: §6.5 · p. 8
• Haptics improved first-turn correctness, and participants said it helped localization, but height cues were not intuitive and required more familiarization.The reported height-cue difficulty concerned frequency-based vertical guidance.Source: §6.5 · p. 8

7 DISCUSSION

• Across both studies, haptic feedback was associated with faster performance and higher confidence, and the conclusion also reports greater behavioral efficiency.Differences in effect strength between studies could reflect sample size or setting, which may have affected attention allocation.Source: §7 · p. 9, §7.2 · p. 9
• Participants described easier localization and intuitive horizontal guidance; their behavior showed fewer unnecessary turns and overshoots, and some called the game fun and creative.These behavioral patterns indicated greater decisiveness and contributed to shorter search times.Source: §7 · p. 9
• With combined horizontal amplitude and vertical frequency cues, performance did not differ across heights, but the study cannot isolate frequency’s contribution.Horizontal cues alone could potentially explain some outcomes.Source: §7 · p. 9
• Vertical cues were perceived as less effective than horizontal guidance, possibly because height relied on a monaural learned pitch–height association rather than bilateral comparison.Some participants found height cues difficult to notice and relied more on vision for vertical judgments.Source: §7 · p. 9

7.1 Future Directions and Limitations

• Future work should examine long-term learning of spatial and multidimensional mappings and interactions with crossmodal correspondences.Other proposed directions include encoding signal identity and integrating multiple simultaneous haptic channels.Source: §7.1 · p. 9
• Practical deployment remains to be assessed in naturalistic settings with clutter, competing stimuli, and varied target velocities; depth encoding and spatial-audio combinations were not evaluated.Participants had brief exposure, and VR tasks do not fully represent real-world complexity.Source: §7.1 · p. 9
• Because the study compared continuous guidance only with no haptics, it cannot establish whether feedback must continue after initial orientation.The authors propose comparing continuous guidance with brief initial cues and feedback that stops when the target enters view.Source: §7.1 · p. 9
• Further evaluation should use larger samples to assess stability and compare horizontal-only cues, workload, and varied frequency encodings.Observed power in Experiment 2 was approximately .60, though significant effects were detected.Source: §7.1 · p. 9