SuperSuit: An Isomorphic Bimodal Interface for Scalable Mobile Manipulation

Tongqing Chen; Hang Wu; Jiasen Wang; Xiaotao Li

cs.ROMar 6, 2026

SuperSuit: An Isomorphic Bimodal Interface for Scalable Mobile Manipulation

Authors: Tongqing Chen, Hang Wu, Jiasen Wang, Xiaotao Li

Organizations: Tsinghua University, Beijing, China. · TARS Robotics, Shanghai, China.

Abstract

Long-horizon mobile manipulation needs whole-body demonstrations, and collecting them is the bottleneck. Keeping the robot in the loop ties the dataset to robot uptime, while recording human motion without it instead requires a conversion step that decides the joint-space action the policy learns. SuperSuit is a single wearable that both teleoperates a wheeled bimanual robot and, with the robot absent, records demonstrations for it. Its arms are mechanically aligned to the robot's, so recorded joint motion needs no inverse kinematics and no retargeting; one body-worn tracker drives the torso and the planar base through a deadband that keeps posture micro-sway out of the base command. Both roles emit the same 22-D command over the torso, planar base, two 7-DoF arms, and two grippers, with arm targets expressed as forward-looking increments of the observed joint state. On three real-robot tasks this lets active demonstrations replace most robot-in-the-loop collection without changing policy quality. With a 10-episode robot-grounded seed held fixed, the remaining 100 episodes can be recorded without the robot, whereas active-only training falls well short. SuperSuit moves the limit on dataset size from the robot to the operator.

Figures & tables

Figure 1: The SuperSuit Framework. Our untethered wearable interface translates human embodiment to whole-body robot control via isomorphic arm manipulation and deadbanded base locomotion. Both teleoperation and active collection produce demonstrations for imitation learning.

Figure 2: System Architecture of SuperSuit. The wearable signals are separated into a mechanical-arm stream for isomorphic arm mapping and a tracker stream for torso and base control. The tracker pose is converted to torso targets and fixed-frame planar-base velocities, followed by a velocity-level deadband. The commands drive the robot during teleoperation and are logged in both collection modes for the offline annotation pipeline, where an LLM combines Paraformer transcriptions with kinematic events to propose subtask boundaries that a human annotator verifies and corrects.

Figure 3: Remote Teleoperation. The operator drives the robot in real time with SuperSuit in (a) Pick-and-Place, (b) Blocks Collection, and (c) Crate Stacking.

Figure 4: Active Demonstration. Robot-free collection under the same whole-body action schema.

Modality	Pick-Place	Blocks	Stacking
BRS-style [ 16 ]	56.8	20.7	11.4
SuperSuit (Teleop)	65.1	24.3	13.2
SuperSuit (Active)	151.4	51.9	32.5

Table 1: Collection throughput (episodes/hour). BRS-style denotes our in-house baseline.

	Policy Success Rate (%)			Effective Throughput (success/hour)
Training Data Composition	Pick-Place	Blocks	Stacking	Pick-Place	Blocks	Stacking
Teleop 10 / Active 0	10%	0%	0%	4.2	0.0	0.0
Teleop 110 / Active 0	85%	65%	40%	41.9	11.4	3.8
Teleop 10 / Active 100	85%	60%	40%	88.8	23.3	9.5

Table 2: Policy success rate and effective throughput under different training data compositions.

Table 8

Architecture	Pick-Place	Blocks	Stacking
ACT [ 40 ]	70%	30%	20%
$\pi_{0.5}$ [ 2 ]	85%	60%	40%
$\pi_{0.5}^{+}$	85%	65%	50%

Table 5: Ablation on policy backbone and subtask annotations.

Appendix figures & tables11 assets

Supplementary material from the paper’s appendix.

Appendix

Component	Specification	Role in SuperSuit experiments
Left/right arms	7 DoF per arm	Bimanual manipulation and object transport
Left/right grippers	1 DoF parallel gripper per arm	Binary or continuous grasping command
Active torso	Lift, yaw, and pitch	Aligns upper-body workspace with the task scene
Mobile base	Omnidirectional $v_{x}$ , $v_{y}$ , $\omega_{z}$ command	Planar locomotion during long-horizon tasks
Global camera	Chest-mounted RGB camera, $640\times 512$ at 30 fps	Global scene observation
Wrist cameras	Left/right fisheye RGB cameras, $640\times 512$ at 30 fps	Local manipulation observation near each gripper

Appendix

Table 6: Robot hardware summary.

Panel of Figure 7 — Figure 7: Two-finger gripper pair. The SuperSuit-side gripper used for data collection and the matching robot-side gripper, which share finger geometry and wrist-camera placement.

Subsystem	Implementation	Notes
Arm interface	Isomorphic left/right wearable arms	Robot-aligned joints
Joint modules	Dynamixel XL330-M288-T servos	Joint logging
Gripper interface	Robot-matched gripper handles	Matched appearance
Spatial tracker	HTC Vive Tracker on a back-fixed bracket, with HTC Base Station 2.0	Base/torso tracking
Audio input	Headset microphone	Narration capture
Wearable mass	3.2 kg	Full wearable system

Appendix

Table 7: Wearable hardware summary.

Parameter	Meaning	Value
$K_{x}$	Forward velocity scale	$1.0$
$K_{y}$	Lateral velocity scale	$1.0$
$K_{\omega}$	Yaw velocity scale	$1.0$
$\epsilon_{x}$	Forward deadband threshold	$0.03$ m/s
$\epsilon_{y}$	Lateral deadband threshold	$0.03$ m/s
$\epsilon_{\omega}$	Yaw deadband threshold	$0.05$ rad/s

Appendix

Table 8: Base-command parameters. Values are conservative engineering defaults for indoor mobile manipulation and can be adjusted for a specific base.

Field	Content	Dimension or format
Chest RGB	Global scene image	$640\times 512\times 3$ at 30 fps
Left wrist RGB	Left fisheye local manipulation image	$640\times 512\times 3$ at 30 fps
Right wrist RGB	Right fisheye local manipulation image	$640\times 512\times 3$ at 30 fps
Raw proprioception	$[\text{left-arm},\text{left-gripper},\text{right-arm},\text{right-gripper},\text{torso}]$	19-D vector
Action	$[\text{left-arm},\text{left-gripper},\text{right-arm},\text{right-gripper},\text{torso},\text{base}]$	22-D vector
Audio	Headset microphone stream	Timestamped WAV file

Appendix

Table 9: Raw collected data record.

Figure 8: Subtask annotation interface. The annotator presents the task-level subtask vocabulary, batch pre-labeling controls, synchronized playback, timeline markers, and multi-view robot observations. Human annotators use this interface to verify model-proposed subtask boundaries and correct the language label associated with each interval.

Parameter	ACT	$\pi_{0.5}$	$\pi_{0.5}^{+}$
Training steps	$50{,}000$	$10{,}000$	$10{,}000$
Batch size	$256$	$256$	$256$
Optimizer	AdamW	AdamW	AdamW
Learning rate	$1\times 10^{-4}$	$5\times 10^{-5}$	$5\times 10^{-5}$
Language input	None	Task instruction	Task instruction
Auxiliary language supervision	None	None	Subtask annotations

Appendix

Table 10: Policy training hyperparameters.

Task	Randomization	Success criterion	Time Limit
Pick-and-Place	Object and box poses	Target objects placed in box	50 s
Blocks Collection	Block pose/shape; receptacle pose	Required blocks collected	180 s
Crate Stacking	Crate and target poses	Crate stably stacked	100 s

Appendix

Table 11: Task-level evaluation protocol.

Figure 9: Model inference rollout for Task 1. The sequence shows the inferred Pick-and-Place behavior, where the robot approaches the tabletop scene, grasps the objects, and places them into the box.

Figure 10: Model inference rollout for Task 3. The sequence shows the inferred Crate Stacking behavior, where the robot approaches the source crate, lifts and transports it to the target stack, places it, and retracts after completion.

Condition	Pick-Place	Blocks	Stacking
Teleop 0 / Active 110	7 [18.1,56.7]	2 [2.8,30.1]	1 [0.9,23.6]
Teleop 10 / Active 100	17 [64.0,94.8]	12 [38.7,78.1]	8 [21.9,61.3]
Teleop 110 / Active 0	17 [64.0,94.8]	13 [43.3,81.9]	8 [21.9,61.3]
BRS-style	16 [58.4,91.9]	13 [43.3,81.9]	7 [18.1,56.7]

Appendix

Table 12: Best-checkpoint outcomes (successes/20) and 95% Wilson intervals.

Explore similar work

SkillWeave: Weaving Heterogeneous Demonstrations into Long-Horizon Manipulation Skills
Oct 8, 2026Ryosei Tamura, Xiaoxiang Dong, Uksang Yoo +4Robotic Manipulation Long-Horizon Robotic Manipulation
Robotics Institute, Carnegie Mellon University, USA. · Department of Computer Science, Keio University, Japan. · College of Engineering, University of California, Berkeley, USA. +1
HoMMI: Learning Whole-Body Mobile Manipulation from Human Demonstrations
Mar 3, 2026Xiaomeng Xu, Jisang Park, Han Zhang +6Mobile Manipulation Whole-Body Robotic Manipulation
Stanford University · Toyota Research Institute
Mobile UMI: Cross-View Diffusion Policy with Decoupled Kinematics for Mobile Manipulation
May 20, 2026Haoran Huang, Haonan Dong, Huixu DongLong-Horizon Robotic Manipulation Mobile Manipulation
Zhejiang University, Hangzhou 310027, China

Oct 8, 2026J/K move · Enter open · S save

SkillWeave: Weaving Heterogeneous Demonstrations into Long-Horizon Manipulation Skills

Ryosei Tamura, Xiaoxiang Dong, Uksang Yoo, Yuemin Mao+3

Robotics Institute, Carnegie Mellon University, USA. · Department of Computer Science, Keio University, Japan. · College of Engineering, University of California, Berkeley, USA.+1

Robotic Manipulation Long-Horizon Robotic Manipulation Robot Imitation Learning Contact-Rich Robotic Manipulation Diffusion Policy

Open paper AskPDF arXiv

Dexterous manipulation requires both large-scale task progression and precise contact-rich interaction, making it challenging to collect demonstrations that effectively support both regimes. We present SkillWeave, a heterogeneous demonstration framework for long-horizon dexterous manipulation that combines teleoperation for coarse reaching and transport with kinesthetic teaching for precise, contact-rich skills. To address the visual mismatch introduced by the demonstrator's presence during kinesthetic data collection, we propose an object-mask-conditioned diffusion policy that uses offline object segmentation for training supervision and a lightweight learned mask predictor at deployment, avoiding online segmentation and image inpainting. To mitigate distribution shift between independently trained sub-task policies, we introduce successor-aware terminal steering, which selects among actions sampled from the predecessor policy to guide the system toward states supported by the successor's demonstrated initial-state distribution. Across three real-world long-horizon tasks, SkillWeave achieves 27% average end-to-end success. Mask-conditioned kinesthetic policies improve dexterous sub-task success to an average of 65%, while successor-aware handoffs achieve an average composition efficiency of 87%. These results show that matching demonstration modality to interaction regime, explicitly addressing kinesthetic visual mismatch, and steering policy handoffs toward successor-supported states substantially improves long-horizon dexterous manipulation. Videos and code are available at skillweave-authors.github.io .

Abstract

Explore similar work

SkillWeave: Weaving Heterogeneous Demonstrations into Long-Horizon Manipulation Skills

HoMMI: Learning Whole-Body Mobile Manipulation from Human Demonstrations

Mobile UMI: Cross-View Diffusion Policy with Decoupled Kinematics for Mobile Manipulation