Art/atk: A research platform for assessing and mitigating the sim-to-real gap in robotics...

A study on the use of simulation in synthesizing path-following control policies for autonomous ground robots

H Zhang, S Caldararu, A Young, A Ruiz… - arXiv preprint arXiv …, 2024 - arxiv.org

We report results obtained and insights gained while answering the following question: how
effective is it to use a simulator to establish path following control policies for an autonomous …

Quantifying the sim2real gap for GPS and IMU sensors

I Mahajan, H Unjhawala, H Zhang, Z Zhou… - arXiv preprint arXiv …, 2024 - arxiv.org

Simulation can and should play a critical role in the development and testing of algorithms
for autonomous agents. What might reduce its impact is the``sim2real''gap--the algorithm …

CUPS: Improving Human Pose-Shape Estimators with Conformalized Deep Uncertainty

H Zhang, L Carlone - arXiv preprint arXiv:2412.10431, 2024 - arxiv.org

We introduce CUPS, a novel method for learning sequence-to-sequence 3D human shapes
and poses from RGB videos with uncertainty quantification. To improve on top of prior work …

Self-Supervised Learning of Dynamic Planar Manipulation of Free-End Cables

J Wang, H Huang, V Lim, H Zhang, J Ichnowski… - arXiv preprint arXiv …, 2024 - arxiv.org

Dynamic manipulation of free-end cables has applications for cable management in homes,
warehouses and manufacturing plants. We present a supervised learning approach for …

CHAMP: Conformalized 3D Human Multi-Hypothesis Pose Estimators

H Zhang, L Carlone - arXiv preprint arXiv:2407.06141, 2024 - arxiv.org

We introduce CHAMP, a novel method for learning sequence-to-sequence, multi-hypothesis
3D human poses from 2D keypoints by leveraging a conditional distribution with a diffusion …

WeightedPose: Generalizable Cross-Pose Estimation via Weighted SVD

X Cheng, H Yu, H Zhang, W Deng - arXiv preprint arXiv:2405.02241, 2024 - arxiv.org

We present a novel method for robotic manipulation tasks in human environments that
require reasoning about the 3D geometric relationship between a pair of objects. Traditional …

Theoretical Analysis for Expectation-Maximization-Based Multi-Model 3D Registration

D Jin, H Zhang, K Chang - arXiv preprint arXiv:2405.08991, 2024 - arxiv.org

We perform detailed theoretical analysis of an expectation-maximization-based algorithm
recently proposed in for solving a variation of the 3D registration problem, named multi …

Learning Correspondence for Deformable Objects

P Sundaresan, A Ganapathi, H Zhang… - arXiv preprint arXiv …, 2024 - arxiv.org

We investigate the problem of pixelwise correspondence for deformable objects, namely
cloth and rope, by comparing both classical and learning-based methods. We choose cloth …

Towards Improving Learning from Demonstration Algorithms via MCMC Methods

C Qi, E Sun, H Zhang - arXiv preprint arXiv:2405.02243, 2024 - arxiv.org

Behavioral cloning, or more broadly, learning from demonstrations (LfD) is a priomising
direction for robot policy learning in complex scenarios. Albeit being straightforward to …