Guanya Shi

Michael O'Connell, Guanya Shi, Xichen Shi et al.

Executing safe and precise flight maneuvers in dynamic high-speed winds is important for the ongoing commoditization of uninhabited aerial vehicles (UAVs). However, because the relationship between various wind conditions and its effect on aircraft maneuverability is not well understood, it is challenging to design effective robot controllers using traditional control design methods. We present Neural-Fly, a learning-based approach that allows rapid online adaptation by incorporating pretrained representations through deep learning. Neural-Fly builds on two key observations that aerodynamics in different wind conditions share a common representation and that the wind-specific part lies in a low-dimensional space. To that end, Neural-Fly uses a proposed learning algorithm, domain adversarially invariant meta-learning (DAIML), to learn the shared representation, only using 12 minutes of flight data. With the learned representation as a basis, Neural-Fly then uses a composite adaptation law to update a set of linear coefficients for mixing the basis elements. When evaluated under challenging wind conditions generated with the Caltech Real Weather Wind Tunnel, with wind speeds up to 43.6 kilometers/hour (12.1 meters/second), Neural-Fly achieves precise flight control with substantially smaller tracking error than state-of-the-art nonlinear and adaptive controllers. In addition to strong empirical performance, the exponential stability of Neural-Fly results in robustness guarantees. Last, our control design extrapolates to unseen wind conditions, is shown to be effective for outdoor flights with only onboard sensors, and can transfer across drones with minimal performance degradation.

Siheng Zhao, Yuanhang Zhang, Ziqi Lu et al.

Humanoid robots hold great promise for operating in human-centered environments, yet ladder climbing remains one of the most challenging tasks due to sparse footholds and handholds, complex whole-body coordination, and sensitivity to perception and control errors. We present \textbf{LadderMan}, a unified system that enables humanoid robots to robustly climb diverse ladders and perform manipulation under such constrained conditions. Our climbing policy is built on a scalable two-stage learning pipeline, where we use hybrid motion tracking to learn multiple climbing experts from a single reference motion, and distill these experts into a unified depth-based visuomotor climbing policy via hybrid imitation and reinforcement learning. To enable real-world deployment, we leverage vision foundation models to bridge the sim-to-real gap in depth perception. Building on the learned climbing policy, we further train a separate manipulation policy using a dual-agent formulation, allowing stable on-ladder manipulation via teleoperation. Experiments demonstrate that LadderMan achieves robust ladder climbing across a wide range of geometries, successfully transfers to real-world hardware in a zero-shot manner, and supports various manipulation tasks under challenging ladder constraints. Video results are available at https://ladderman-robot.github.io .

Yanjie Ze, Siheng Zhao, Weizhuo Wang et al.

Large-scale data has driven breakthroughs in robotics, from language models to vision-language-action models in bimanual manipulation. However, humanoid robotics lacks equally effective data collection frameworks. Existing humanoid teleoperation systems either use decoupled control or depend on expensive motion capture setups. We introduce TWIST2, a portable, mocap-free humanoid teleoperation and data collection system that preserves full whole-body control while advancing scalability. Our system leverages PICO4U VR for obtaining real-time whole-body human motions, with a custom 2-DoF robot neck (cost around $250) for egocentric vision, enabling holistic human-to-humanoid control. We demonstrate long-horizon dexterous and mobile humanoid skills and we can collect 100 demonstrations in 15 minutes with an almost 100% success rate. Building on this pipeline, we propose a hierarchical visuomotor policy framework that autonomously controls the full humanoid body based on egocentric vision. Our visuomotor policy successfully demonstrates whole-body dexterous manipulation and dynamic kicking tasks. The entire system is fully reproducible and open-sourced at https://yanjieze.com/TWIST2 . Our collected dataset is also open-sourced at https://twist-data.github.io .

Younggyo Seo, Carmelo Sferrazza, Juyue Chen et al.

Massively parallel simulation has reduced reinforcement learning (RL) training time for robots from days to minutes. However, achieving fast and reliable sim-to-real RL for humanoid control remains difficult due to the challenges introduced by factors such as high dimensionality and domain randomization. In this work, we introduce a simple and practical recipe based on off-policy RL algorithms, i.e., FastSAC and FastTD3, that enables rapid training of humanoid locomotion policies in just 15 minutes with a single RTX 4090 GPU. Our simple recipe stabilizes off-policy RL algorithms at massive scale with thousands of parallel environments through carefully tuned design choices and minimalist reward functions. We demonstrate rapid end-to-end learning of humanoid locomotion controllers on Unitree G1 and Booster T1 robots under strong domain randomization, e.g., randomized dynamics, rough terrain, and push perturbations, as well as fast training of whole-body human-motion tracking policies. We provide videos and open-source implementation at: https://younggyo.me/fastsac-humanoid.

Zhen Wu, Xiaoyu Huang, Lujie Yang et al.

While recent advances in humanoid locomotion have achieved stable walking on varied terrains, capturing the agility and adaptivity of highly dynamic human motions remains an open challenge. In particular, agile parkour in complex environments demands not only low-level robustness, but also human-like motion expressiveness, long-horizon skill composition, and perception-driven decision-making. In this paper, we present Perceptive Humanoid Parkour (PHP), a modular framework that enables humanoid robots to autonomously perform long-horizon, vision-based parkour across challenging obstacle courses. Our approach first leverages motion matching, formulated as nearest-neighbor search in a feature space, to compose retargeted atomic human skills into long-horizon kinematic trajectories. This framework enables the flexible composition and smooth transition of complex skill chains while preserving the elegance and fluidity of dynamic human motions. Next, we train motion-tracking reinforcement learning (RL) expert policies for these composed motions, and distill them into a single depth-based, multi-skill student policy, using a combination of DAgger and RL. Crucially, the combination of perception and skill composition enables autonomous, context-aware decision-making: using only onboard depth sensing and a discrete 2D velocity command, the robot selects and executes whether to step over, climb onto, vault or roll off obstacles of varying geometries and heights. We validate our framework with extensive real-world experiments on a Unitree G1 humanoid robot, demonstrating highly dynamic parkour skills such as climbing tall obstacles up to 1.25m (96% robot height), as well as long-horizon multi-obstacle traversal with closed-loop adaptation to real-time obstacle perturbations.

Andrew Wagenmaker, Guanya Shi, Kevin Jamieson

Learning to control unknown nonlinear dynamical systems is a fundamental problem in reinforcement learning and control theory. A commonly applied approach is to first explore the environment (exploration), learn an accurate model of it (system identification), and then compute an optimal controller with the minimum cost on this estimated system (policy optimization). While existing work has shown that it is possible to learn a uniformly good model of the system~\citep{mania2020active}, in practice, if we aim to learn a good controller with a low cost on the actual system, certain system parameters may be significantly more critical than others, and we therefore ought to focus our exploration on learning such parameters. In this work, we consider the setting of nonlinear dynamical systems and seek to formally quantify, in such settings, (a) which parameters are most relevant to learning a good controller, and (b) how we can best explore so as to minimize uncertainty in such parameters. Inspired by recent work in linear systems~\citep{wagenmaker2021task}, we show that minimizing the controller loss in nonlinear systems translates to estimating the system parameters in a particular, task-dependent metric. Motivated by this, we develop an algorithm able to efficiently explore the system to reduce uncertainty in this metric, and prove a lower bound showing that our approach learns a controller at a near-instance-optimal rate. Our algorithm relies on a general reduction from policy optimization to optimal experiment design in arbitrary systems, and may be of independent interest. We conclude with experiments demonstrating the effectiveness of our method in realistic nonlinear robotic systems.

Brent Yi, Hongsuk Choi, Himanshu Gaurav Singh et al.

Likelihood-based policy gradient methods are the dominant approach for training robot control policies from rewards. These methods rely on differentiable action likelihoods, which constrain policy outputs to simple distributions like Gaussians. In this work, we show how flow matching policy gradients -- a recent framework that bypasses likelihood computation -- can be made effective for training and fine-tuning more expressive policies in challenging robot control settings. We introduce an improved objective that enables success in legged locomotion, humanoid motion tracking, and manipulation tasks, as well as robust sim-to-real transfer on two humanoid robots. We then present ablations and analysis on training dynamics. Results show how policies can exploit the flow representation for exploration when training from scratch, as well as improved fine-tuning robustness over baselines.

Yifang Chen, Yingbing Huang, Simon S. Du et al.

Representation learning based on multi-task pretraining has become a powerful approach in many domains. In particular, task-aware representation learning aims to learn an optimal representation for a specific target task by sampling data from a set of source tasks, while task-agnostic representation learning seeks to learn a universal representation for a class of tasks. In this paper, we propose a general and versatile algorithmic and theoretic framework for \textit{active representation learning}, where the learner optimally chooses which source tasks to sample from. This framework, along with a tractable meta algorithm, allows most arbitrary target and source task spaces (from discrete to continuous), covers both task-aware and task-agnostic settings, and is compatible with deep representation learning practices. We provide several instantiations under this framework, from bilinear and feature-based nonlinear to general nonlinear cases. In the bilinear case, by leveraging the non-uniform spectrum of the task representation and the calibrated source-target relevance, we prove that the sample complexity to achieve $\varepsilon$-excess risk on target scales with $ (k^*)^2 \|v^*\|_2^2 \varepsilon^{-2}$ where $k^*$ is the effective dimension of the target and $\|v^*\|_2^2 \in (0,1]$ represents the connection between source and target space. Compared to the passive one, this can save up to $\frac{1}{d_W}$ of sample complexity, where $d_W$ is the task space dimension. Finally, we demonstrate different instantiations of our meta algorithm in synthetic datasets and robotics problems, from pendulum simulations to real-world drone flight datasets. On average, our algorithms outperform baselines by $20\%-70\%$.

Wenli Xiao, Tairan He, John Dolan et al.

A critical goal of autonomy and artificial intelligence is enabling autonomous robots to rapidly adapt in dynamic and uncertain environments. Classic adaptive control and safe control provide stability and safety guarantees but are limited to specific system classes. In contrast, policy adaptation based on reinforcement learning (RL) offers versatility and generalizability but presents safety and robustness challenges. We propose SafeDPA, a novel RL and control framework that simultaneously tackles the problems of policy adaptation and safe reinforcement learning. SafeDPA jointly learns adaptive policy and dynamics models in simulation, predicts environment configurations, and fine-tunes dynamics models with few-shot real-world data. A safety filter based on the Control Barrier Function (CBF) on top of the RL policy is introduced to ensure safety during real-world deployment. We provide theoretical safety guarantees of SafeDPA and show the robustness of SafeDPA against learning errors and extra perturbations. Comprehensive experiments on (1) classic control problems (Inverted Pendulum), (2) simulation benchmarks (Safety Gym), and (3) a real-world agile robotics platform (RC Car) demonstrate great superiority of SafeDPA in both safety and task performance, over state-of-the-art baselines. Particularly, SafeDPA demonstrates notable generalizability, achieving a 300% increase in safety rate compared to the baselines, under unseen disturbances in real-world experiments.

Pablo Ortega-Kral, Eliot Xing, Arthur Bucker et al.

Despite recent efforts to collect multi-task, multi-embodiment datasets, to design recipes for training Vision-Language-Action models (VLAs), and to showcase these models on different robot platforms, generalist cross-embodiment robot capabilities remains a largely elusive ideal. Progress is limited by fragmented infrastructure: most robot code is highly specific to the exact setup the user decided on, which adds major overhead when attempting to reuse, recycle, or share artifacts between users. We present RIO (Robot I/O), an open source Python framework that provides flexible, lightweight components for robot control, teleoperation, data formatting, sensor configuration, and policy deployment across diverse hardware platforms and morphologies. RIO provides abstractions that enable users to make any choice and to switch between them, with minimal reconfiguration effort. We validate RIO on VLA deployment workflows across three morphologies (single-arm, bimanual, humanoid) and four hardware platforms with varying grippers and cameras. Using teleoperated data collected with RIO, we fine-tune state-of-the-art VLAs including $π_{0.5}$ and GR00T on household tasks such as pick-and-place, folding, and bowl scrubbing. By open sourcing all our efforts, we hope the community can accelerate their pace of robot learning on real-world robot hardware. Additional details at: https://robot-i-o.github.io

Max Fu, Justin Yu, Karim El-Refai et al.

"Code-as-Policy" considers how executable code can complement data-intensive Vision-Language-Action (VLA) methods, yet their effectiveness as autonomous controllers for embodied manipulation remains underexplored. We present CaP-X, an open-access framework for systematically studying Code-as-Policy agents in robot manipulation. At its core is CaP-Gym, an interactive environment in which agents control robots by synthesizing and executing programs that compose perception and control primitives. Building on this foundation, CaP-Bench evaluates frontier language and vision-language models across varying levels of abstraction, interaction, and perceptual grounding. Across 12 models, CaP-Bench reveals a consistent trend: performance improves with human-crafted abstractions but degrades as these priors are removed, exposing a dependence on designer scaffolding. At the same time, we observe that this gap can be mitigated through scaling agentic test-time computation--through multi-turn interaction, structured execution feedback, visual differencing, automatic skill synthesis, and ensembled reasoning--substantially improves robustness even when agents operate over low-level primitives. These findings allow us to derive CaP-Agent0, a training-free framework that recovers human-level reliability on several manipulation tasks in simulation and on real embodiments. We further introduce CaP-RL, showing reinforcement learning with verifiable rewards improves success rates and transfers from sim2real with minimal gap. Together, CaP-X provides a principled, open-access platform for advancing embodied coding agents.

Kevin Huang, Rwik Rana, Alexander Spitzer et al.

Precise arbitrary trajectory tracking for quadrotors is challenging due to unknown nonlinear dynamics, trajectory infeasibility, and actuation limits. To tackle these challenges, we present Deep Adaptive Trajectory Tracking (DATT), a learning-based approach that can precisely track arbitrary, potentially infeasible trajectories in the presence of large disturbances in the real world. DATT builds on a novel feedforward-feedback-adaptive control structure trained in simulation using reinforcement learning. When deployed on real hardware, DATT is augmented with a disturbance estimator using L1 adaptive control in closed-loop, without any fine-tuning. DATT significantly outperforms competitive adaptive nonlinear and model predictive controllers for both feasible smooth and infeasible trajectories in unsteady wind fields, including challenging scenarios where baselines completely fail. Moreover, DATT can efficiently run online with an inference time less than 3.2 ms, less than 1/4 of the adaptive nonlinear model predictive control baseline

Jacob Sacks, Rwik Rana, Kevin Huang et al.

A major challenge in robotics is to design robust policies which enable complex and agile behaviors in the real world. On one end of the spectrum, we have model-free reinforcement learning (MFRL), which is incredibly flexible and general but often results in brittle policies. In contrast, model predictive control (MPC) continually re-plans at each time step to remain robust to perturbations and model inaccuracies. However, despite its real-world successes, MPC often under-performs the optimal strategy. This is due to model quality, myopic behavior from short planning horizons, and approximations due to computational constraints. And even with a perfect model and enough compute, MPC can get stuck in bad local optima, depending heavily on the quality of the optimization algorithm. To this end, we propose Deep Model Predictive Optimization (DMPO), which learns the inner-loop of an MPC optimization algorithm directly via experience, specifically tailored to the needs of the control problem. We evaluate DMPO on a real quadrotor agile trajectory tracking task, on which it improves performance over a baseline MPC algorithm for a given computational budget. It can outperform the best MPC algorithm by up to 27% with fewer samples and an end-to-end policy trained with MFRL by 19%. Moreover, because DMPO requires fewer samples, it can also achieve these benefits with 4.3X less memory. When we subject the quadrotor to turbulent wind fields with an attached drag plate, DMPO can adapt zero-shot while still outperforming all baselines. Additional results can be found at https://tinyurl.com/mr2ywmnw.

Wenli Xiao, Haotian Lin, Andy Peng et al.

Supervised fine-tuning (SFT) has become the de facto post-training strategy for large vision-language-action (VLA) models, but its reliance on costly human demonstrations limits scalability and generalization. We propose Probe, Learn, Distill (PLD), a three-stage plug-and-play framework that improves VLAs through residual reinforcement learning (RL) and distribution-aware data collection. In Stage 1, we train lightweight residual actors to probe failure regions of the VLA generalist. In Stage 2, we use a hybrid rollout scheme that aligns collected trajectories with the generalist's deployment distribution while capturing recovery behaviors. In Stage 3, we distill the curated trajectories back into the generalist with standard SFT. PLD achieves near-saturated 99% task success on LIBERO, over 50% gains in SimplerEnv, and 100% success on real-world Franka and YAM arm manipulation tasks. Ablations show that residual probing and distribution-aware replay are key to collecting deployment-aligned data that improves both seen and unseen tasks, offering a scalable path toward self-improving VLA models.

Chaoyi Pan, Changhao Wang, Haozhi Qi et al.

Learning dexterous and agile policy for humanoid and dexterous hand control requires large-scale demonstrations, but collecting robot-specific data is prohibitively expensive. In contrast, abundant human motion data is readily available from motion capture, videos, and virtual reality, which could help address the data scarcity problem. However, due to the embodiment gap and missing dynamic information like force and torque, these demonstrations cannot be directly executed on robots. To bridge this gap, we propose Scalable Physics-Informed DExterous Retargeting (SPIDER), a physics-based retargeting framework to transform and augment kinematic-only human demonstrations to dynamically feasible robot trajectories at scale. Our key insight is that human demonstrations should provide global task structure and objective, while large-scale physics-based sampling with curriculum-style virtual contact guidance should refine trajectories to ensure dynamical feasibility and correct contact sequences. SPIDER scales across diverse 9 humanoid/dexterous hand embodiments and 6 datasets, improving success rates by 18% compared to standard sampling, while being 10X faster than reinforcement learning (RL) baselines, and enabling the generation of a 2.4M frames dynamic-feasible robot dataset for policy learning. As a universal physics-based retargeting method, SPIDER can work with diverse quality data and generate diverse and high-quality data to enable efficient policy learning with methods like RL.

Cunxi Dai, Haoran Chang, Aditya Nisal et al.

Behavior cloning for contact-rich bimanual manipulation remains challenging because diverse demonstrations are expensive to collect, and even small disturbances can push the system into off-manifold states where no recovery supervision is available. We propose PGDG, a data generation framework with zero-shot curation that expands a single demonstration into a compact dataset of physically plausible, successful, and diverse recovery behaviors without additional human labeling. PGDG iterates between a physics-grounded sampler and a dataset curator, where the curator selects informative, non-redundant, and recoverable behaviors to update the sampling distribution toward under-covered recovery modes, and the sampler draws physically plausible rollout candidates from this updated distribution and retains successful trajectories. To further improve data quality, PGDG applies short-horizon sampling-based control to relabel selected risky states with corrective actions. Across four bimanual manipulation tasks, PGDG consistently outperforms spatial-only augmentation in both simulation and zero-shot real-world transfer. On RotateBox-Pitch, success improves from 38% to 93% in simulation and from 35% to 82% in the real world. PGDG also enables effective foundation models fine-tuning such as GR00T, increasing success from 46% to 77%. Additional results are available in our website: https://cunxid.github.io/PGDG/.

Haoru Xue, Tairan He, Zi Wang et al.

Recent progress in GPU-accelerated, photorealistic simulation has opened a scalable data-generation path for robot learning, where massive physics and visual randomization allow policies to generalize beyond curated environments. Building on these advances, we develop a teacher-student-bootstrap learning framework for vision-based humanoid loco-manipulation, using articulated-object interaction as a representative high-difficulty benchmark. Our approach introduces a staged-reset exploration strategy that stabilizes long-horizon privileged-policy training, and a GRPO-based fine-tuning procedure that mitigates partial observability and improves closed-loop consistency in sim-to-real RL. Trained entirely on simulation data, the resulting policy achieves robust zero-shot performance across diverse door types and outperforms human teleoperators by up to 31.7% in task completion time under the same whole-body control stack. This represents the first humanoid sim-to-real policy capable of diverse articulated loco-manipulation using pure RGB perception.

Chaoyi Pan, Giri Anantharaman, Nai-Chieh Huang et al.

Generative models, like flows and diffusions, have recently emerged as popular and efficacious policy parameterizations in robotics. There has been much speculation as to the factors underlying their successes, ranging from capturing multi-modal action distribution to expressing more complex behaviors. In this work, we perform a comprehensive evaluation of popular generative control policies (GCPs) on common behavior cloning (BC) benchmarks. We find that GCPs do not owe their success to their ability to capture multi-modality or to express more complex observation-to-action mappings. Instead, we find that their advantage stems from iterative computation, as long as intermediate steps are supervised during training and this supervision is paired with a suitable level of stochasticity. As a validation of our findings, we show that a minimum iterative policy (MIP), a lightweight two-step regression-based policy, essentially matches the performance of flow GCPs, and often outperforms distilled shortcut models. Our results suggest that the distribution-fitting component of GCPs is less salient than commonly believed, and point toward new design spaces focusing solely on control performance. Project page: https://simchowitzlabpublic.github.io/much-ado-about-noising-project/

Harsh Gupta, Xiaofeng Guo, Huy Ha et al.

We introduce UMI-on-Air, a framework for embodiment-aware deployment of embodiment-agnostic manipulation policies. Our approach leverages diverse, unconstrained human demonstrations collected with a handheld gripper (UMI) to train generalizable visuomotor policies. A central challenge in transferring these policies to constrained robotic embodiments-such as aerial manipulators-is the mismatch in control and robot dynamics, which often leads to out-of-distribution behaviors and poor execution. To address this, we propose Embodiment-Aware Diffusion Policy (EADP), which couples a high-level UMI policy with a low-level embodiment-specific controller at inference time. By integrating gradient feedback from the controller's tracking cost into the diffusion sampling process, our method steers trajectory generation towards dynamically feasible modes tailored to the deployment embodiment. This enables plug-and-play, embodiment-aware trajectory adaptation at test time. We validate our approach on multiple long-horizon and high-precision aerial manipulation tasks, showing improved success rates, efficiency, and robustness under disturbances compared to unguided diffusion baselines. Finally, we demonstrate deployment in previously unseen environments, using UMI demonstrations collected in the wild, highlighting a practical pathway for scaling generalizable manipulation skills across diverse-and even highly constrained-embodiments. All code, data, checkpoints, and result videos can be found at umi-on-air.github.io.

Zhanpeng Luo, Ce Zhang, Silong Yong et al.

Multi-modal Large Language Models (MLLMs) have demonstrated strong capabilities in general-purpose perception and reasoning, but they still struggle with tasks that require spatial understanding of the 3D world. To address this, we introduce pySpatial, a visual programming framework that equips MLLMs with the ability to interface with spatial tools via Python code generation. Given an image sequence and a natural-language query, the model composes function calls to spatial tools including 3D reconstruction, camera-pose recovery, novel-view rendering, etc. These operations convert raw 2D inputs into an explorable 3D scene, enabling MLLMs to reason explicitly over structured spatial representations. Notably, pySpatial requires no gradient-based fine-tuning and operates in a fully zero-shot setting. Experimental evaluations on the challenging MindCube and Omni3D-Bench benchmarks demonstrate that our framework pySpatial consistently surpasses strong MLLM baselines; for instance, it outperforms GPT-4.1-mini by 12.94% on MindCube. Furthermore, we conduct real-world indoor navigation experiments where the robot can successfully traverse complex environments using route plans generated by pySpatial, highlighting the practical effectiveness of our approach.

Zeji Yi, Chaoyi Pan, Guanya Shi et al.

Sampling-based optimization (SBO), like cross-entropy method and evolutionary algorithms, has achieved many successes in solving non-convex problems without gradients, yet its convergence is poorly understood. In this paper, we establish a non-asymptotic convergence analysis for SBO through the lens of smoothing. Specifically, we recast SBO as gradient descent on a smoothed objective, mirroring noise-conditioned score ascent in diffusion models. Our first contribution is a landscape analysis of the smoothed objective, demonstrating how smoothing helps escape local minima and uncovering a fundamental coverage-optimality trade-off: smoothing renders the landscape more benign by enlarging the locally convex region around the global minimizer, but at the cost of introducing an optimality gap. Building on this insight, we establish non-asymptotic convergence guarantees for SBO algorithms to a neighborhood of the global minimizer. Furthermore, we propose an annealed SBO algorithm, Diffusion-Inspired Dual-Annealing (DIDA), which is provably convergent to the global optimum. We conduct extensive numerical experiments to verify our landscape results and also demonstrate the compelling performance of DIDA compared to other gradient-free optimization methods. Lastly, we discuss implications of our results for diffusion models.

Peiqi Yu, Philip Huang, Chaitanya Chawla et al.

Robotic assembly systems traditionally require substantial manual engineering effort to integrate new tasks, adapt to new environments, and improve performance over time. This paper presents a framework for autonomous integration and continuous improvement of robotic assembly systems based on Skill Graph representations. A Skill Graph organizes robot capabilities as verb-based skills, explicitly linking semantic descriptions (verbs and nouns) with executable policies, pre-conditions, post-conditions, and evaluators. We show how Skill Graphs enable rapid system integration by supporting semantic-level planning over skills, while simultaneously grounding execution through well-defined interfaces to robot controllers and perception modules. After initial deployment, the same Skill Graph structure supports systematic data collection and closed-loop performance improvement, enabling iterative refinement of skills and their composition. We demonstrate how this approach unifies system configuration, execution, evaluation, and learning within a single representation, providing a scalable pathway toward adaptive and reusable robotic assembly systems. The code is at https://github.com/intelligent-control-lab/AIDF.

Tairan He, Zhengyi Luo, Wenli Xiao et al. · cmu

We present Human to Humanoid (H2O), a reinforcement learning (RL) based framework that enables real-time whole-body teleoperation of a full-sized humanoid robot with only an RGB camera. To create a large-scale retargeted motion dataset of human movements for humanoid robots, we propose a scalable "sim-to-data" process to filter and pick feasible motions using a privileged motion imitator. Afterwards, we train a robust real-time humanoid motion imitator in simulation using these refined motions and transfer it to the real humanoid robot in a zero-shot manner. We successfully achieve teleoperation of dynamic whole-body motions in real-world scenarios, including walking, back jumping, kicking, turning, waving, pushing, boxing, etc. To the best of our knowledge, this is the first demonstration to achieve learning-based real-time whole-body humanoid teleoperation.

Tairan He, Chong Zhang, Wenli Xiao et al.

Legged robots navigating cluttered environments must be jointly agile for efficient task execution and safe to avoid collisions with obstacles or humans. Existing studies either develop conservative controllers (< 1.0 m/s) to ensure safety, or focus on agility without considering potentially fatal collisions. This paper introduces Agile But Safe (ABS), a learning-based control framework that enables agile and collision-free locomotion for quadrupedal robots. ABS involves an agile policy to execute agile motor skills amidst obstacles and a recovery policy to prevent failures, collaboratively achieving high-speed and collision-free navigation. The policy switch in ABS is governed by a learned control-theoretic reach-avoid value network, which also guides the recovery policy as an objective function, thereby safeguarding the robot in a closed loop. The training process involves the learning of the agile policy, the reach-avoid value network, the recovery policy, and an exteroception representation network, all in simulation. These trained modules can be directly deployed in the real world with onboard sensing and computation, leading to high-speed and collision-free navigation in confined indoor and outdoor spaces with both static and dynamic obstacles.

Tairan He, Jiawei Gao, Wenli Xiao et al.

Humanoid robots hold the potential for unparalleled versatility in performing human-like, whole-body skills. However, achieving agile and coordinated whole-body motions remains a significant challenge due to the dynamics mismatch between simulation and the real world. Existing approaches, such as system identification (SysID) and domain randomization (DR) methods, often rely on labor-intensive parameter tuning or result in overly conservative policies that sacrifice agility. In this paper, we present ASAP (Aligning Simulation and Real-World Physics), a two-stage framework designed to tackle the dynamics mismatch and enable agile humanoid whole-body skills. In the first stage, we pre-train motion tracking policies in simulation using retargeted human motion data. In the second stage, we deploy the policies in the real world and collect real-world data to train a delta (residual) action model that compensates for the dynamics mismatch. Then, ASAP fine-tunes pre-trained policies with the delta action model integrated into the simulator to align effectively with real-world dynamics. We evaluate ASAP across three transfer scenarios: IsaacGym to IsaacSim, IsaacGym to Genesis, and IsaacGym to the real-world Unitree G1 humanoid robot. Our approach significantly improves agility and whole-body coordination across various dynamic motions, reducing tracking error compared to SysID, DR, and delta dynamics learning baselines. ASAP enables highly agile motions that were previously difficult to achieve, demonstrating the potential of delta action learning in bridging simulation and real-world dynamics. These results suggest a promising sim-to-real direction for developing more expressive and agile humanoids.

Ri-Zhao Qiu, Shiqi Yang, Xuxin Cheng et al.

Training manipulation policies for humanoid robots with diverse data enhances their robustness and generalization across tasks and platforms. However, learning solely from robot demonstrations is labor-intensive, requiring expensive tele-operated data collection which is difficult to scale. This paper investigates a more scalable data source, egocentric human demonstrations, to serve as cross-embodiment training data for robot learning. We mitigate the embodiment gap between humanoids and humans from both the data and modeling perspectives. We collect an egocentric task-oriented dataset (PH2D) that is directly aligned with humanoid manipulation demonstrations. We then train a human-humanoid behavior policy, which we term Human Action Transformer (HAT). The state-action space of HAT is unified for both humans and humanoid robots and can be differentiably retargeted to robot actions. Co-trained with smaller-scale robot data, HAT directly models humanoid robots and humans as different embodiments without additional supervision. We show that human data improves both generalization and robustness of HAT with significantly better data collection efficiency. Code and data: https://human-as-robot.github.io/

Zeji Yi, Chaoyi Pan, Guanqi He et al.

Sampling-based Model Predictive Control (MPC) has been a practical and effective approach in many domains, notably model-based reinforcement learning, thanks to its flexibility and parallelizability. Despite its appealing empirical performance, the theoretical understanding, particularly in terms of convergence analysis and hyperparameter tuning, remains absent. In this paper, we characterize the convergence property of a widely used sampling-based MPC method, Model Predictive Path Integral Control (MPPI). We show that MPPI enjoys at least linear convergence rates when the optimization is quadratic, which covers time-varying LQR systems. We then extend to more general nonlinear systems. Our theoretical analysis directly leads to a novel sampling-based MPC algorithm, CoVariance-Optimal MPC (CoVo-MPC) that optimally schedules the sampling covariance to optimize the convergence rate. Empirically, CoVo-MPC significantly outperforms standard MPPI by 43-54% in both simulations and real-world quadrotor agile control tasks. Videos and Appendices are available at \url{https://lecar-lab.github.io/CoVO-MPC/}.

Haoru Xue, Xiaoyu Huang, Dantong Niu et al.

Vision-language-action (VLA) models have demonstrated strong semantic understanding and zero-shot generalization, yet most existing systems assume an accurate low-level controller with hand-crafted action "vocabulary" such as end-effector pose or root velocity. This assumption confines prior work to quasi-static tasks and precludes the agile, whole-body behaviors required by humanoid whole-body control (WBC) tasks. To capture this gap in the literature, we start by introducing the first sim-to-real-ready, vision-language, closed-loop benchmark for humanoid WBC, comprising over 150 tasks from 10 categories. We then propose LeVERB: Latent Vision-Language-Encoded Robot Behavior, a hierarchical latent instruction-following framework for humanoid vision-language WBC, the first of its kind. At the top level, a vision-language policy learns a latent action vocabulary from synthetically rendered kinematic demonstrations; at the low level, a reinforcement-learned WBC policy consumes these latent verbs to generate dynamics-level commands. In our benchmark, LeVERB can zero-shot attain a 80% success rate on simple visual navigation tasks, and 58.5% success rate overall, outperforming naive hierarchical whole-body VLA implementation by 7.8 times.

Haotian Lin, Pengcheng Wang, Jeff Schneider et al.

Model-based reinforcement learning algorithms that combine model-based planning and learned value/policy prior have gained significant recognition for their high data efficiency and superior performance in continuous control. However, we discover that existing methods that rely on standard SAC-style policy iteration for value learning, directly using data generated by the planner, often result in \emph{persistent value overestimation}. Through theoretical analysis and experiments, we argue that this issue is deeply rooted in the structural policy mismatch between the data generation policy that is always bootstrapped by the planner and the learned policy prior. To mitigate such a mismatch in a minimalist way, we propose a policy regularization term reducing out-of-distribution (OOD) queries, thereby improving value learning. Our method involves minimum changes on top of existing frameworks and requires no additional computation. Extensive experiments demonstrate that the proposed approach improves performance over baselines such as TD-MPC2 by large margins, particularly in 61-DoF humanoid tasks. View qualitative results at https://darthutopian.github.io/tdmpc_square/.

Lujie Yang, Xiaoyu Huang, Zhen Wu et al.

A dominant paradigm for teaching humanoid robots complex skills is to retarget human motions as kinematic references to train reinforcement learning (RL) policies. However, existing retargeting pipelines often struggle with the significant embodiment gap between humans and robots, producing physically implausible artifacts like foot-skating and penetration. More importantly, common retargeting methods neglect the rich human-object and human-environment interactions essential for expressive locomotion and loco-manipulation. To address this, we introduce OmniRetarget, an interaction-preserving data generation engine based on an interaction mesh that explicitly models and preserves the crucial spatial and contact relationships between an agent, the terrain, and manipulated objects. By minimizing the Laplacian deformation between the human and robot meshes while enforcing kinematic constraints, OmniRetarget generates kinematically feasible trajectories. Moreover, preserving task-relevant interactions enables efficient data augmentation, from a single demonstration to different robot embodiments, terrains, and object configurations. We comprehensively evaluate OmniRetarget by retargeting motions from OMOMO, LAFAN1, and our in-house MoCap datasets, generating over 8-hour trajectories that achieve better kinematic constraint satisfaction and contact preservation than widely used baselines. Such high-quality data enables proprioceptive RL policies to successfully execute long-horizon (up to 30 seconds) parkour and loco-manipulation skills on a Unitree G1 humanoid, trained with only 5 reward terms and simple domain randomization shared by all tasks, without any learning curriculum.

Yichao Zhong, Chong Zhang, Tairan He et al.

Real-world legged locomotion systems often need to reconcile agility and safety for different scenarios. Moreover, the underlying dynamics are often unknown and time-variant (e.g., payload, friction). In this paper, we introduce BAS (Bridging Adaptivity and Safety), which builds upon the pipeline of prior work Agile But Safe (ABS)(He et al.) and is designed to provide adaptive safety even in dynamic environments with uncertainties. BAS involves an agile policy to avoid obstacles rapidly and a recovery policy to prevent collisions, a physical parameter estimator that is concurrently trained with agile policy, and a learned control-theoretic RA (reach-avoid) value network that governs the policy switch. Also, the agile policy and RA network are both conditioned on physical parameters to make them adaptive. To mitigate the distribution shift issue, we further introduce an on-policy fine-tuning phase for the estimator to enhance its robustness and accuracy. The simulation results show that BAS achieves 50% better safety than baselines in dynamic environments while maintaining a higher speed on average. In real-world experiments, BAS shows its capability in complex environments with unknown physics (e.g., slippery floors with unknown frictions, unknown payloads up to 8kg), while baselines lack adaptivity, leading to collisions or. degraded agility. As a result, BAS achieves a 19.8% increase in speed and gets a 2.36 times lower collision rate than ABS in the real world. Videos: https://adaptive-safe-locomotion.github.io.

Siheng Zhao, Yanjie Ze, Yue Wang et al.

Humanoid whole-body loco-manipulation promises transformative capabilities for daily service and warehouse tasks. While recent advances in general motion tracking (GMT) have enabled humanoids to reproduce diverse human motions, these policies lack the precision and object awareness required for loco-manipulation. To this end, we introduce ResMimic, a two-stage residual learning framework for precise and expressive humanoid control from human motion data. First, a GMT policy, trained on large-scale human-only motion, serves as a task-agnostic base for generating human-like whole-body movements. An efficient but precise residual policy is then learned to refine the GMT outputs to improve locomotion and incorporate object interaction. To further facilitate efficient training, we design (i) a point-cloud-based object tracking reward for smoother optimization, (ii) a contact reward that encourages accurate humanoid body-object interactions, and (iii) a curriculum-based virtual object controller to stabilize early training. We evaluate ResMimic in both simulation and on a real Unitree G1 humanoid. Results show substantial gains in task success, training efficiency, and robustness over strong baselines. Videos are available at https://resmimic.github.io/ .

Lars Ankile, Zhenyu Jiang, Rocky Duan et al.

Recent advances in behavior cloning (BC) have enabled impressive visuomotor control policies. However, these approaches are limited by the quality of human demonstrations, the manual effort required for data collection, and the diminishing returns from offline data. In comparison, reinforcement learning (RL) trains an agent through autonomous interaction with the environment and has shown remarkable success in various domains. Still, training RL policies directly on real-world robots remains challenging due to sample inefficiency, safety concerns, and the difficulty of learning from sparse rewards for long-horizon tasks, especially for high-degree-of-freedom (DoF) systems. We present a recipe that combines the benefits of BC and RL through a residual learning framework. Our approach leverages BC policies as black-box bases and learns lightweight per-step residual corrections via sample-efficient off-policy RL. We demonstrate that our method requires only sparse binary reward signals and can effectively improve manipulation policies on high-degree-of-freedom (DoF) systems in both simulation and the real world. In particular, we demonstrate, to the best of our knowledge, the first successful real-world RL training on a humanoid robot with dexterous hands. Our results demonstrate state-of-the-art performance in various vision-based tasks, pointing towards a practical pathway for deploying RL in the real world.

Guo Ning Sue, Yogita Choudhary, Richard Desatnik et al.

Ensuring safety via safety filters in real-world robotics presents significant challenges, particularly when the system dynamics is complex or unavailable. To handle this issue, learning-based safety filters recently gained popularity, which can be classified as model-based and model-free methods. Existing model-based approaches requires various assumptions on system model (e.g., control-affine), which limits their application in complex systems, and existing model-free approaches need substantial modifications to standard RL algorithms and lack versatility. This paper proposes a simple, plugin-and-play, and effective model-free safety filter learning framework. We introduce a novel reward formulation and use Q-learning to learn Q-value functions to safeguard arbitrary task specific nominal policies via filtering out their potentially unsafe actions. The threshold used in the filtering process is supported by our theoretical analysis. Due to its model-free nature and simplicity, our framework can be seamlessly integrated with various RL algorithms. We validate the proposed approach through simulations on double integrator and Dubin's car systems and demonstrate its effectiveness in real-world experiments with a soft robotic limb.

Tairan He, Zhengyi Luo, Xialin He et al.

We present OmniH2O (Omni Human-to-Humanoid), a learning-based system for whole-body humanoid teleoperation and autonomy. Using kinematic pose as a universal control interface, OmniH2O enables various ways for a human to control a full-sized humanoid with dexterous hands, including using real-time teleoperation through VR headset, verbal instruction, and RGB camera. OmniH2O also enables full autonomy by learning from teleoperated demonstrations or integrating with frontier models such as GPT-4. OmniH2O demonstrates versatility and dexterity in various real-world whole-body tasks through teleoperation or autonomy, such as playing multiple sports, moving and manipulating objects, and interacting with humans. We develop an RL-based sim-to-real pipeline, which involves large-scale retargeting and augmentation of human motion datasets, learning a real-world deployable policy with sparse sensor input by imitating a privileged teacher policy, and reward designs to enhance robustness and stability. We release the first humanoid whole-body control dataset, OmniH2O-6, containing six everyday tasks, and demonstrate humanoid whole-body skill learning from teleoperated datasets.

Weici Pan, Guanya Shi, Yiheng Lin et al.

We study a variant of online optimization in which the learner receives $k$-round $\textit{delayed feedback}$ about hitting cost and there is a multi-step nonlinear switching cost, i.e., costs depend on multiple previous actions in a nonlinear manner. Our main result shows that a novel Iterative Regularized Online Balanced Descent (iROBD) algorithm has a constant, dimension-free competitive ratio that is $O(L^{2k})$, where $L$ is the Lipschitz constant of the switching cost. Additionally, we provide lower bounds that illustrate the Lipschitz condition is required and the dependencies on $k$ and $L$ are tight. Finally, via reductions, we show that this setting is closely related to online control problems with delay, nonlinear dynamics, and adversarial disturbances, where iROBD directly offers constant-competitive online policies.

Guanya Shi, Kamyar Azizzadenesheli, Michael O'Connell et al.

We present an online multi-task learning approach for adaptive nonlinear control, which we call Online Meta-Adaptive Control (OMAC). The goal is to control a nonlinear system subject to adversarial disturbance and unknown $\textit{environment-dependent}$ nonlinear dynamics, under the assumption that the environment-dependent dynamics can be well captured with some shared representation. Our approach is motivated by robot control, where a robotic system encounters a sequence of new environmental conditions that it must quickly adapt to. A key emphasis is to integrate online representation learning with established methods from control theory, in order to arrive at a unified framework that yields both control-theoretic and learning-theoretic guarantees. We provide instantiations of our approach under varying conditions, leading to the first non-asymptotic end-to-end convergence guarantee for multi-task nonlinear control. OMAC can also be integrated with deep representation learning. Experiments show that OMAC significantly outperforms conventional adaptive control approaches which do not learn the shared representation, in inverted pendulum and 6-DoF drone control tasks under varying wind conditions.

Michael O'Connell, Guanya Shi, Xichen Shi et al.

Realtime model learning proves challenging for complex dynamical systems, such as drones flying in variable wind conditions. Machine learning technique such as deep neural networks have high representation power but is often too slow to update onboard. On the other hand, adaptive control relies on simple linear parameter models can update as fast as the feedback control loop. We propose an online composite adaptation method that treats outputs from a deep neural network as a set of basis functions capable of representing different wind conditions. To help with training, meta-learning techniques are used to optimize the network output useful for adaptation. We validate our approach by flying a drone in an open air wind tunnel under varying wind conditions and along challenging trajectories. We compare the result with other adaptive controller with different basis function sets and show improvement over tracking and prediction errors.

Guanya Shi, Wolfgang Hönig, Xichen Shi et al.

We present Neural-Swarm2, a learning-based method for motion planning and control that allows heterogeneous multirotors in a swarm to safely fly in close proximity. Such operation for drones is challenging due to complex aerodynamic interaction forces, such as downwash generated by nearby drones and ground effect. Conventional planning and control methods neglect capturing these interaction forces, resulting in sparse swarm configuration during flight. Our approach combines a physics-based nominal dynamics model with learned Deep Neural Networks (DNNs) with strong Lipschitz properties. We make use of two techniques to accurately predict the aerodynamic interactions between heterogeneous multirotors: i) spectral normalization for stability and generalization guarantees of unseen data and ii) heterogeneous deep sets for supporting any number of heterogeneous neighbors in a permutation-invariant manner without reducing expressiveness. The learned residual dynamics benefit both the proposed interaction-aware multi-robot motion planning and the nonlinear tracking control design because the learned interaction forces reduce the modelling errors. Experimental results demonstrate that Neural-Swarm2 is able to generalize to larger swarms beyond training cases and significantly outperforms a baseline nonlinear tracking controller with up to three times reduction in worst-case tracking errors.

Guanya Shi, Yifeng Zhu, Jonathan Tremblay et al.

Deep learning-based object pose estimators are often unreliable and overconfident especially when the input image is outside the training domain, for instance, with sim2real transfer. Efficient and robust uncertainty quantification (UQ) in pose estimators is critically needed in many robotic tasks. In this work, we propose a simple, efficient, and plug-and-play UQ method for 6-DoF object pose estimation. We ensemble 2-3 pre-trained models with different neural network architectures and/or training data sources, and compute their average pairwise disagreement against one another to obtain the uncertainty quantification. We propose four disagreement metrics, including a learned metric, and show that the average distance (ADD) is the best learning-free metric and it is only slightly worse than the learned metric, which requires labeled target data. Our method has several advantages compared to the prior art: 1) our method does not require any modification of the training process or the model inputs; and 2) it needs only one forward pass for each model. We evaluate the proposed UQ method on three tasks where our uncertainty quantification yields much stronger correlations with pose estimation errors than the baselines. Moreover, in a real robot grasping task, our method increases the grasping success rate from 35% to 90%.

Yashwanth Kumar Nakka, Anqi Liu, Guanya Shi et al.

Learning-based control algorithms require data collection with abundant supervision for training. Safe exploration algorithms ensure the safety of this data collection process even when only partial knowledge is available. We present a new approach for optimal motion planning with safe exploration that integrates chance-constrained stochastic optimal control with dynamics learning and feedback control. We derive an iterative convex optimization algorithm that solves an \underline{Info}rmation-cost \underline{S}tochastic \underline{N}onlinear \underline{O}ptimal \underline{C}ontrol problem (Info-SNOC). The optimization objective encodes control cost for performance and exploration cost for learning, and the safety is incorporated as distributionally robust chance constraints. The dynamics are predicted from a robust regression model that is learned from data. The Info-SNOC algorithm is used to compute a sub-optimal pool of safe motion plans that aid in exploration for learning unknown residual dynamics under safety constraints. A stable feedback controller is used to execute the motion plan and collect data for model learning. We prove the safety of rollout from our exploration method and reduction in uncertainty over epochs, thereby guaranteeing the consistency of our learning method. We validate the effectiveness of Info-SNOC by designing and implementing a pool of safe trajectories for a planar robot. We demonstrate that our approach has higher success rate in ensuring safety when compared to a deterministic trajectory optimization approach.

Guanya Shi, Wolfgang Hönig, Yisong Yue et al.

In this paper, we present Neural-Swarm, a nonlinear decentralized stable controller for close-proximity flight of multirotor swarms. Close-proximity control is challenging due to the complex aerodynamic interaction effects between multirotors, such as downwash from higher vehicles to lower ones. Conventional methods often fail to properly capture these interaction effects, resulting in controllers that must maintain large safety distances between vehicles, and thus are not capable of close-proximity flight. Our approach combines a nominal dynamics model with a regularized permutation-invariant Deep Neural Network (DNN) that accurately learns the high-order multi-vehicle interactions. We design a stable nonlinear tracking controller using the learned model. Experimental results demonstrate that the proposed controller significantly outperforms a baseline nonlinear tracking controller with up to four times smaller worst-case height tracking errors. We also empirically demonstrate the ability of our learned model to generalize to larger swarm sizes.

Guanya Shi, Yiheng Lin, Soon-Jo Chung et al.

This paper presents competitive algorithms for a novel class of online optimization problems with memory. We consider a setting where the learner seeks to minimize the sum of a hitting cost and a switching cost that depends on the previous $p$ decisions. This setting generalizes Smoothed Online Convex Optimization. The proposed approach, Optimistic Regularized Online Balanced Descent, achieves a constant, dimension-free competitive ratio. Further, we show a connection between online optimization with memory and online control with adversarial disturbances. This connection, in turn, leads to a new constant-competitive policy for a rich class of online control problems.

Anqi Liu, Guanya Shi, Soon-Jo Chung et al.

We study the problem of safe learning and exploration in sequential control problems. The goal is to safely collect data samples from operating in an environment, in order to learn to achieve a challenging control goal (e.g., an agile maneuver close to a boundary). A central challenge in this setting is how to quantify uncertainty in order to choose provably-safe actions that allow us to collect informative data and reduce uncertainty, thereby achieving both improved controller safety and optimality. To address this challenge, we present a deep robust regression model that is trained to directly predict the uncertainty bounds for safe exploration. We derive generalization bounds for learning, and connect them with safety and stability bounds in control. We demonstrate empirically that our robust regression approach can outperform the conventional Gaussian process (GP) based safe exploration in settings where it is difficult to specify a good GP prior.

Guanya Shi, Xichen Shi, Michael O'Connell et al.

Precise near-ground trajectory control is difficult for multi-rotor drones, due to the complex aerodynamic effects caused by interactions between multi-rotor airflow and the environment. Conventional control methods often fail to properly account for these complex effects and fall short in accomplishing smooth landing. In this paper, we present a novel deep-learning-based robust nonlinear controller (Neural Lander) that improves control performance of a quadrotor during landing. Our approach combines a nominal dynamics model with a Deep Neural Network (DNN) that learns high-order interactions. We apply spectral normalization (SN) to constrain the Lipschitz constant of the DNN. Leveraging this Lipschitz property, we design a nonlinear feedback linearization controller using the learned model and prove system stability with disturbance rejection. To the best of our knowledge, this is the first DNN-based nonlinear feedback controller with stability guarantees that can utilize arbitrarily large neural nets. Experimental results demonstrate that the proposed controller significantly outperforms a Baseline Nonlinear Tracking Controller in both landing and cross-table trajectory tracking cases. We also empirically show that the DNN generalizes well to unseen data outside the training domain.