MIRROR: Visual Motion Imitation via Real-time Retargeting and Teleoperation with Parallel Differential Inverse Kinematics

Junheng Li; Lizhi Yang; Aaron D. Ames

cs.ROMar 25, 2026

MIRROR: Visual Motion Imitation via Real-time Retargeting and Teleoperation with Parallel Differential Inverse Kinematics

Authors: Junheng Li, Lizhi Yang, Aaron D. Ames

Organizations: Department of Mechanical and Civil Engineering, California Institute of Technology, Pasadena, USA.

Abstract

Real-time humanoid teleoperation requires inverse kinematics (IK) solvers that are both responsive and constraint-safe under kinematic redundancy and self-collision constraints. While differential IK enables efficient online retargeting, its locally linearized updates are inherently basin-dependent and often become trapped near joint limits, singularities, or active collision boundaries, leading to unsafe or stagnant behavior. We propose a GPU-parallelized, continuation-based differential IK that improves escape from such constraint-induced local minima while preserving real-time performance, promoting safety and stability. Multiple constrained IK quadratic programs are evaluated in parallel, together with a self-collision avoidance control barrier function (CBF), and a Lyapunov-based progression criterion selects updates that reduce the final global task-space error. The method is paired with a visual skeletal pose estimation pipeline that enables robust, real-time upper-body teleoperation on the THEMIS humanoid robot hardware in real-world tasks.

Figures & tables

Fig. 1 : Real-time human-humanoid motion mirroring teleoperation. Full experiment videos https://youtu.be/LvwIwOTTu9g

Fig. 3 : Snapshots of real-time teleoperation. Pose mirroring and mimicking.

Fig. 4 : Snapshots of useful real-world tasks through MIRROR teleoperation.

Method	Batch $K$	$\bm{\eta}$	Solve Time [ms] $\downarrow$	$\bm{\|\|\bar{e}_{\textrm{hand,l}}\|\|}$ [mm] $\downarrow$	$\bm{\|\|\bar{e}_{\textrm{hand,l}}^{\textrm{final}}\|\|}$ [mm] $\downarrow$	Self Collision $\downarrow$	Singularity $\downarrow$	Stagnation $\downarrow$
Global IK (SQP)	1	–	$176.57\pm 4.20$	$12.44\pm 3.12$	$6.58\pm 2.23$	$42/250$	$0/250$	$0/250$
Monolithic QP	1	–	$0.58\pm 0.05$	$18.57\pm 9.82$	$16.23\pm 4.12$	$135/250$	$17/250$	$44/250$
Distributed QP	1	–	$0.94\pm 0.06$	$17.72\pm 10.68$	$10.44\pm 4.56$	$142/250$	$19/250$	$46/250$
Parallel Dist. QPs - Continuation $\checkmark$ - $V$ certificate $\times$	$256$	–	$4.28\pm 0.14$	$16.61\pm 7.95$	$11.45\pm 3.34$	$68/250$	$9/250$	$45/250$
	$4096$	–	$4.08\pm 0.06$	$16.61\pm 7.95$	$11.45\pm 3.34$	$68/250$	$9/250$	$45/250$
Parallel Mono. QPs - Continuation $\checkmark$ - $V$ certificate $\checkmark$	$256$	0.0005	$3.94\pm 0.09$	$16.61\pm 7.86$	$10.38\pm 4.20$	$52/250$	$9/250$	$45/250$

TABLE I : Ablation study on differential IK solving methods over long-horizon hand tracking.

Fig. 5 : Comparison of motion capture devices tracking hand location.

Fig. 6 : Latency breakdown with different hardware setups. Only the main components of each module are presented here; overhead and communication latencies are only counted in the ”Module Total” latencies.

Fig. 7 : Left hand tracking performance on hardware. Hardware is using an optimization-based whole-body controller to execute joint-level commands and balancing tasks.

Inverse Kinematics

Inverse kinematics maps desired robot or articulated-body motions to joint configurations while handling constraints, redundancy, and multiple solutions.

See all 39 papers

MimicIK: Real-Time Generative Inverse Kinematics from Teleoperation with FK Consistency
Jun 13, 2026Jiahao Yang, Shenhao Yan, Fan Feng +5Robotic Manipulation Inverse Kinematics
1Ising AI · 2CUHK-Shenzhen · Ising AI
HJCD-IK: GPU-Accelerated Inverse Kinematics through Batched Hybrid Jacobian Coordinate Descent
Oct 8, 2025Cael Yasutake, Andrew H. Liu, Zachary Kingston +1Inverse Kinematics Robot Kinematics
Columbia University, NY, USA. · Purdue University, IN, USA. · Barnard College, Columbia University, NY, USA, & Dartmouth College, NH, USA.
Kilohertz-Safe: A Scalable Framework for Constrained Dexterous Retargeting
Mar 31, 2026Yinxiao Tian, Ziyi Yang, Zinan Zhao +1Motion Retargeting Robotic Control
Department of Automation at the University of Science and Technology of China, Hefei, Anhui, China, 230026.

Jun 13, 2026J/K move · Enter open · S save

MimicIK: Real-Time Generative Inverse Kinematics from Teleoperation with FK Consistency

Jiahao Yang, Shenhao Yan, Fan Feng, Chengsi Yao+4

1Ising AI · 2CUHK-Shenzhen · Ising AI

Robotic Manipulation Inverse Kinematics Robot Kinematics

Open paper AskPDF arXiv

Inverse kinematics (IK) remains a critical bottleneck for real-time robot manipulation. Classical numerical solvers achieve high geometric precision but often suffer from discontinuous branch switching and unstable behavior near kinematic singularities during closed-loop deployment. Meanwhile, learned IK approaches frequently struggle to balance spatial accuracy, motion smoothness, and real-time efficiency, particularly when trained on noisy human teleoperation data. We present MimicIK, a real-time generative inverse kinematics framework that learns smooth and robust joint-space motion priors from teleoperation demonstrations through conditional flow matching. Given the current joint configuration and a target end-effector pose, MimicIK predicts continuous delta-joint commands using an efficient two-step iterative refinement process based on a Minimal Iterative Policy (MIP) backbone. To enforce physical consistency, we further introduce an FK consistency loss, a differentiable forward-kinematics regularization that penalizes task-space deviations from the target pose during training. We evaluate MimicIK on a real-world 6-DOF robot dataset containing 8,848 teleoperation demonstrations. MimicIK achieves a mean position error of 4.65 mm, a 10 mm success rate of 92.01%, and a trajectory spike rate of only 7.99%. Compared with a UNet diffusion baseline, our method improves both spatial accuracy and motion smoothness while reducing inference latency from 21.66 ms to 6.74 ms. Furthermore, unlike deterministic MLP baselines that catastrophically diverge under out-of-distribution deployment, MimicIK remains stable near singular configurations and enables robust 20 Hz real-time control on deployment hardware.

Abstract

Inverse Kinematics

Explore similar work

MimicIK: Real-Time Generative Inverse Kinematics from Teleoperation with FK Consistency

HJCD-IK: GPU-Accelerated Inverse Kinematics through Batched Hybrid Jacobian Coordinate Descent

Kilohertz-Safe: A Scalable Framework for Constrained Dexterous Retargeting