Pei Lin

OmniCalib, a target-free workflow that calibrates the full upper limbs---all 14 arm joint zeros and the extrinsics of both wrist and chest cameras---as well as lower limbs and the multi-camera head rig, and writes only supported corrections to the CAD model.

Continuous reliability weights for foot kinematics can slash trajectory errors by over 83% in dynamic locomotion scenarios.

Calibration without external systems can cut foot-height errors in humanoid robots by over 50%, enhancing kinematic accuracy in real-world applications.

Achieving over 97% success in long-horizon robot manipulation, TemporalFlow-VLA reveals the critical role of execution history in action prediction.

V-Link bridges the gap between visual perception and action control, boosting robotic manipulation success rates by up to 31.2% through enhanced feature integration.

OVIP-SG outperforms existing frameworks by preserving small object instances and enhancing retrieval accuracy, revolutionizing the mapping of fine-grained objects in 3D environments.

MVUCF achieves a remarkable 98.9% on LIBERO, showcasing a 22.4-point improvement on LIBERO-Plus and a 23.3-point increase in task success rates, all without additional inference costs.

Joint pretraining on Ego2Robot-synthesized data boosts robot generalization, achieving unprecedented scale and diversity in training datasets.

Explicitly incorporating contact priors into visuo-tactile policies can boost manipulation success rates by over 21% in real-world scenarios.

Qwen-RobotNav redefines navigation by allowing real-time reconfiguration of strategies, achieving unprecedented flexibility and performance across diverse tasks.

Qwen-RobotManip achieves a 20% relative improvement over the previous state-of-the-art in robotic manipulation, showcasing unprecedented generalization capabilities from diverse, open-source datasets.

Language-driven video generation in Qwen-RobotWorld achieves unprecedented accuracy in predicting robotic actions, outperforming existing models across key benchmarks.

One model to control them all: Qwen-VLA achieves impressive zero-shot generalization across diverse robotic tasks and embodiments by unifying vision-language-action modeling.