SRL Proxemics: Spatial Guidelines for Supernumerary Robotic Limbs in Near-Body Interactions

TL;DR

This study investigates how users negotiate near-body autonomy for wearable SRLs and introduces SRL Proxemics, a zone- and segment-based framework derived from a Wizard-of-Oz elicitation with 18 participants. Across two autonomy conditions and two near-body tasks, results show that higher autonomy does not automatically enhance safety or trust; instead, users prefer region-specific rules and clear cues, leading to greater perceived safety, trust, and embodiment under participant-defined rules. Physiological arousal concentrates at entry into sensitive zones, and subjective measures consistently favor PDR for assurance, while High-Autonomy can boost perceived intelligence but raise unpredictability. The work proposes a three-tiered zone-responsive control policy and discusses future state-aware overlays to operationalize user-generated proxemics in real SRL controllers, offering actionable guidance for safer, more legible, and trustworthy near-body augmentation.

Abstract

Wearable supernumerary robotic limbs (SRLs) sit at the intersection of human augmentation and embodied AI, transforming into extensions of the human body. However, their movements within the intimate near-body space raise unresolved challenges for perceived safety, user control, and trust. In this paper, we present results from a Wizard-of-Oz study (n=18), where participants completed near-body collaboration tasks with SRLs to explore these challenges. We collected qualitative data through think-aloud protocols and semi-structured interviews, complemented by physiological signals and post-task ratings. Findings indicate that greater autonomy did not inherently enhance perceived safety or trust. Instead, participants identified near-body zones and paired them with clear coordination rules. They also expressed expectations for how different arm components should behave, shaping preferences around autonomy, perceived safety, and trust. Building on these insights, we introduce SRL Proxemics, a zone- and segment-level design framework showing that autonomy is not monolithic: perceived safety hinges on spatially calibrated, legible behaviors, not higher autonomy.

Figures (6)

• Figure 1: Task layouts and wearable SRLs setup. (a) Comfort Zone Task handovers. (b) Control Handover Task: sorting deformable snack packs. (c) Wearable backpack-mounted SRLs with articulated joints and interchangeable end-effectors for near-body manipulation.
• Figure 2: Participant-defined sensitivity zones: orange = "stay an arm’s length away"; yellow = "only if I can stop it instantly"; blue = "close to my body is acceptable here." Labels reflect participants’ desired standoff distances for SRLs entry by body region. Numbers (e.g., 17/18) show how many participants assigned each zone that constraint. (b) Participant-defined segment-level autonomy: Color gradient (yellow to red) reflects reduced comfort with autonomy: yellow for manual/confirm-only control, red for higher autonomy.
• Figure 3: Illustrations of participant-preferred SRLs behaviors. (a--b) Participants preferred SRLs to approach sensitive regions (e.g., head, face) using "pause–move–pause" styles with curved trajectories, rather than direct or unpredictable paths. (c) Autonomous actions were most often assigned to the distal segments (e.g., hands) of the non-dominant-side SRLs to support manipulation while preserving clear roles.
• Figure 4: Preferred SRLs motion strategies near the body. (a–b) Participants favored arc-shaped, whole-arm trajectories over abrupt vertical or frontal intrusions, especially near the face and torso. Lifting motions were more acceptable when the entire limb rose cohesively rather than the hand popping up. (c) Auditory cues like subtle motor hums helped users anticipate SRLs movements occurring behind them.
• Figure 5: Skin conductance responses across four interaction phases under two autonomy conditions (Participant-Defined Rules and High-Autonomy Anchor), and Trust in Automated Systems ratings comparing the same two conditions. (a) SCR across interaction stages. (b) Trust and distrust questionnaire ratings across conditions. * indicates $p < .05$.