Robot Dynamics Learning

Follow this topic Ask about this topic

Robot dynamics learning studies how to predict a robot’s motion and its interactions with objects and environments from states, actions, and sensor data. Methods range from physics-informed models such as Lagrangian and Koopman networks to recurrent models, graph neural networks, Gaussian processes, and learned world models; many combine physical structure with neural residuals or uncertainty estimates. These models support model-predictive control, reinforcement learning, manipulation, and sim-to-real transfer, with evaluation spanning MuJoCo and manipulation benchmarks as well as real arms, legged robots, and soft robots. A central challenge is maintaining accurate, stable long-horizon predictions through contact and changing conditions while keeping models efficient and transferable across tasks and robot morphologies.

Latest papers

90 papers26 papers in the last four weeks, up 86% on the four weeks before. 0.2% of all new papers.

Sep 22, 2026

Manipulation of Deformable Linear Objects Using Model Predictive Path Integral Control with Bidirectional Long Short-Term Memory Learning

The manipulation of Deformable Linear Objects (DLOs) such as cables poses a significant challenge for automation due to their infinite degrees of freedom and non-linear dynamics. In this paper we present a machine learning based optimal control approach for the manipulation of DLOs. This approach is divided into two main components: modeling and control. For modeling the dynamics of the DLO, we propose a learning based approach using a bidirectional Long Short-Term Memory (biLSTM) network. The biLSTM network is trained on synthetic data generated by the MuJoCo physics engine. For manipulating the DLO, a model predictive control strategy that employs Model Predictive Path Integral (MPPI) control is selected. The proposed approach is evaluated through simulation and experiments. The results demonstrate the effectiveness of the proposed method in achieving accurate and efficient manipulation of DLOs.

Lukas Zeh, Johannes Meiwaldt, Zexu Zhou +2

Institute for Control Engineering of Machine Tools and Manufacturing Units, University of Stuttgart, Stuttgart, Germany

Deformable Object Manipulation Robot Dynamics Learning Robotic Manipulation

Sep 22, 2026

History-Conditioned Flow Matching for Probabilistic Dynamics of Tendon-Driven Continuum Robots

Deterministic dynamics modeling of tendon-driven continuum robots remains challenging owing to uncertainties in material behavior, tendon transmission, friction, and contact. Measured joint configurations and nominal tendon commands do not fully characterize these internal mechanical factors, leaving uncertainty in the subsequent motion. We therefore develop a history-conditioned, physics-informed flow-matching framework for probabilistic dynamics prediction, using motion and actuation histories to predict the distribution of the next complete joint configuration. By recursively sampling next-step configurations under prescribed commands, the model predicts distributions of future whole-body motions. In simulation, scenario-specific models achieve five-second trajectory Energy Scores (lower is better) of 12.05 mm under internal friction variation and 9.29 mm under unobserved actuation disturbances. Relative to the conditional variational autoencoder and diffusion baselines, Flow attains lower Energy Scores and coverage closer to the nominal level in both scenarios. Ablations support history and structural conditioning in both scenarios. On the physical robot, predictions under two tendon-command profiles excluded from training capture the principal motion sequences, with five-second Energy Scores of 11.91 and 11.42 mm, lower than the compared baselines. The predicted-to-measured spread ratios are 1.65 and 1.22 (closer to 1 is better). These results support history-conditioned probabilistic dynamics prediction under incomplete mechanical observations.

Hang Yang, Tingcong Liu, Junjie Xiong +2

Flow Matching Robot Dynamics Learning Tendon-Driven Actuation Soft Robotics

Sep 19, 2026

Robot World Models Are Not Invariant to How the Actions Are Written

A robot policy is trained with one of two action parameterizations: absolute joint targets, or deltas relative to the current state. The choice is a live engineering decision in robot learning, and a world model conditioned on actions inherits it silently. We show the inheritance is catastrophic. A latent dynamics model trained on one parameterization and handed the identical commanded trajectory written in the other collapses: retrieval degrades by 2.6-13.4x across three robot datasets and two morphologies, goal-conditioned action selection falls from 53% to 15%, and on PushT the two beliefs about the same future are near-orthogonal (cos = 0.067, worst case -0.377), so the predictor does not degrade gracefully, it answers a different question. This is not a distribution-shift artifact in the usual sense: the two encodings are mutually reconstructible at R^2 = 0.996 given the joint input, so no information is lost, and we give the test that separates a valid re-parameterization from a lossy summary or a sensor swap. The test rejected three of the four axes we proposed. The defect lives in the action channel, which the invariance literature for visual models does not examine: work there concerns crops, jitter and camera pose, while the parameterization of the commands goes unaudited. The repair is averaging over the two encodings, and where it goes matters. Averaging the objective restores task performance by itself; averaging the outputs, safe for probabilities by concavity, is not available for direction-valued prediction, where the normalized mean can score below every member of the orbit. What objective-averaging leaves behind is the tail: worst-case agreement stays at 0.78, a disagreement penalty closes it to 0.995, and over a latent rollout it is the difference between a worst case that erodes and one that holds. On PushT, averaging alone does not repair the axis.

Ahmed Karim, Leon Chlon

Hassana Labs · University College London · University of Oxford

World Models for Robotics Action-Conditioned World Models Robot Dynamics Learning Invariant Representation Learning

Sep 19, 2026

Expert-Play Contouring Control: Faster-than-Demonstration Planning from Slow Expert and Fast Play

Expert demonstrations often specify what a robot should do, but not how fast it can do it. Imitation Learning (IL) inherits demonstration timing, while directly accelerating the learned motion can fail when faster execution changes the robot-object dynamics. We study faster-than-demonstration execution as a dynamics-aware control problem and introduce Expert-Play Contouring Control (EPCC), which combines slow expert demonstrations with fast, non-expert play. Expert demonstrations train a latent trajectory generator whose predictions are reparameterized into a time-independent contour of successful task progression, while play trains a world model (WM) of fast-action outcomes. At deployment, our proposed planner uses the WM to optimize actions that makes maximize progress along the expert-derived contour while penalizing deviation from the intended task evolution. Averaged across three visuomotor manipulation tasks, EPCC achieves a

2.0\times

the throughput of the IL baseline, including

2.2\times

that of the throughput of the strongest acceleration baseline on a task with interaction-sensitive object dynamics. Our analysis shows that the gains concentrate where faster execution changes robot-object evolution. Together, our results highlight a simple yet effective principle: demonstrations provide task intent, while play data provides the dynamic coverage needed to execute that intent faster.

Seunghoon Cho, Wonsuhk Jung, Sundhar Vinodh Sangeetha +1

Seoul National University, Seoul, South Korea · Georgia Institute of Technology, Atlanta, GA, USA

Robot Motion Planning Robotic Manipulation Learning from Demonstration Robot Dynamics Learning

Sep 17, 2026

MetaPusher: Meta Learning and Planning for Nonprehensile Manipulation of Unseen Objects with Rapid Online Adaption

Manipulating previously unseen objects remains challenging, as their dynamics depend on latent physical properties, such as friction and mass distribution, that cannot be inferred from perception alone. Prior experience across objects can provide an initial estimate of unseen object dynamics, but this estimate remains uncertain and can degrade further during sim-to-real transfer. Adapting the dynamics through interaction can progressively refine the estimation, however, updating the model may invalidate the planned trajectory. Successful and efficient manipulation therefore requires both rapid dynamics adaptation and a planning strategy that can incorporate this evolution. In this work, we introduce MetaPusher, a meta-learning and adaptive planning framework for nonprehensile manipulation of unseen objects without prior object-specific interactions. A meta-learned dynamics model rapidly adapts from interactions during task execution, while an adaptive kinodynamic planner updates long-horizon plans by reusing and refining its existing search tree. This coupling enables manipulation and adaptation without a separate data collection phase. We evaluate MetaPusher on unseen objects in simulation and in sim-to-real scenarios, comparing against fine-tuning and active learning methods, MPPI-based control, and a reinforcement learning policy. It achieves lower prediction error and improves task success rate by up to 20%.

Donghyung Lee, Seyedali Golestaneh, Jaskrit Singh +3

Non-Prehensile Robotic Manipulation Robotic Manipulation Robot Learning Robotic Object Pushing Robot Dynamics Learning Long-Horizon Robotic Manipulation+1

Sep 16, 2026

ForwardDLO: Model-Based Bimanual Shape Matching of Unconstrained Deformable Linear Objects

Ropes, cables, and other deformable linear objects appear in tasks from untangling to cable routing and suturing, yet controlling their shape remains a challenge in robot manipulation. We study model-based shape control in a general setting: the object lies unfixated on a support surface and two arms may grasp and move it anywhere along its length. Because each arm chooses a grasp point, direction, and magnitude, the joint action space is combinatorially large, and the dynamics model's per-prediction cost bounds how much of it a planner can search. We present ForwardDLO, a recurrent latent dynamics model for this unfixated bimanual setting that predicts per-segment displacements grounded in the observed rope state at every step. Our model reaches accuracy comparable to more expensive baselines while containing no explicit segment-to-segment operations, which makes batched evaluation of candidate actions cheap. On open-loop prediction of real rope motion it reaches the lowest error of the learned models we evaluate, 13% below the strongest baseline. Within a fixed time budget it scores 8 to 22 times more candidate actions than models of comparable accuracy while matching them in real-world shape matching; and on a simulated routing task at a 30Hz control rate, this throughput converts into 98% task success versus at most 30% for the baselines at their own budgets. We release the model, code, and a dataset of 2.42 million simulated and 14,107 real rope transitions at https://anonymous.4open.science/r/ForwardDLO/

Tim Missal, Berk Guler, Lucas Domingues +3

Technical University of Darmstadt · Honda Research Institute Europe GmbH · School of Electrical and Computer Engineering, Universidade Estadual de Campinas (UNICAMP), Brazil +4

Bimanual Robotic Manipulation Deformable Object Manipulation Robot Dynamics Learning Dynamic Robotic Manipulation Robot Learning Deformable Object Simulation

Sep 15, 2026

Embedding Physics Priors in Robot Learning: A Survey

The rapid progress of artificial intelligence is reshaping robotics and accelerating the adoption of learning-based approaches. While purely data-driven methods have achieved remarkable success in computer vision and natural language processing, robotics remains constrained by limited data, complex real-world interactions, and the need for reliable operation. These challenges have motivated the exploration of physics-embedded robot learning, which embeds physics priors into learning algorithms. By encoding the underlying physical laws and constraints, physics priors can complement limited data with robotics-specific inductive biases, potentially improving generalization, interpretability, and sample efficiency. However, the literature on physics-embedded robot learning remains fragmented across terminology, methodologies, and application domains, making it difficult to assess this growing body of work. This survey reviews physics-embedded robot learning across a broad range of physics priors, robotics applications, and machine learning models, from single-layer perceptrons to generative foundation models. We adopt a unified taxonomy that classifies existing approaches according to their physics embedding: physics-guided inputs, data, and representations; physics-encoded model architectures; and physics-informed training loss functions. Building on this taxonomy, we review methods for robot dynamics learning, trajectory planning, prediction, control, and estimation, together with the corresponding open-source software ecosystem. We identify key open challenges, and outline promising future research directions. Overall, we argue that physics priors provide a particularly relevant robotics-specific inductive bias, complementing rather than replacing data-driven learning, and paving the way toward more generalizable, data-efficient, and trustworthy robotic systems.

Mattia Piccinini, Lucas Schulze, Alice Plebe +10

Professorship of Autonomous Vehicle Systems, Technical University of Munich, 85748 Garching, Germany; Munich Institute of Robotics and Machine Intelligence (MIRMI). · Department of Computer Science, Technical University of Darmstadt, Germany. · Department of Industrial Engineering, University of Trento, 38123 Trento, Italy. +1

Robot Learning Physics-Informed ML Robotic Control Robot Dynamics Learning

Sep 13, 2026

Orientation Control of Soft Robots via Adiabatic Spectral Submanifolds

Soft robots are commonly sought for safety-critical interactions in delicate environments, where accurate position and orientation control is imperative. Model predictive control (MPC) offers a solution, but it requires a model of the robot's infinite-dimensional nonlinear dynamics that is at once accurate and computationally cheap. Recent theory on adiabatic spectral submanifolds (aSSMs) and their applications to soft robots provide data-driven model-reduction methods to construct such models. Here, we extend these methods to identify aSSMs from enlarged observable datasets and upgrade the currently available aSSM-MPC schemes. Evaluated on a high-fidelity finite-element simulation of a pressure-actuated soft arm, our controller reduces position and orientation tracking error by more than 60% compared to existing data-driven baselines.

Aron Karakai, Roshan S. Kaundinya, Mike Yan Michelis +2

Automatic Control Laboratory, ETH Zurich, Switzerland · Institute for Mechanical Systems, ETH Zurich, Switzerland · Soft Robotics Laboratory and the ETH AI Center, ETH Zurich, Switzerland

Robotic Control Soft Robotics Model Predictive Control Robot Trajectory Tracking Robot Dynamics Learning

Sep 6, 2026

Identifying Habit, Physics, and Nuisance in Robot World Models

Teleoperated demonstrations are often multimodal even when the underlying dynamics are nearly deterministic given the executed action. We argue that this multimodality typically mixes three factors--operator habit in action selection, shared physics, and observation nuisance--and that entangled next-observation predictors absorb all three. We formalize the split with a structural causal model a=g(h,z,u), z'=f(z,a), o=r(z,c), and test it with complementary interventions: replacing or shuffling actions at fixed state sharply increases next-state error, whereas appearance and camera changes should not; habit-aware reverse scoring improves ranking of feasible pasts without rewriting the dynamics. The associated adaptation rule is to freeze a shared physics readout and update only a thin interface. On StackCube, DROID, and RH20T this rule improves low-shot transfer relative to training from scratch, retains cleaner dynamics under corrupted adaptation data, and extends from proprioception to pixel observations with multi-view and multi-step checks. We do not equate latent actions with operator habit, and we do not target large-scale video generation benchmarks.

Jinting Hang, Zhenhui Cai

Harvest Praxis

World Models for Robotics Robot Dynamics Learning Data-Efficient Robot Learning

Sep 3, 2026

Modular Deep Recurrent Neural Network: Application to Quadrotors

A modular deep Recurrent Neural Network (RNN) is introduced to facilitate the process of deploying various architectures of RNNs, and to automatically compute derivatives for gradient-based learning methods. The modularity leads to a set of new architectures, one of which includes feedforward inter-layer connections. By adding feedforward inter-layer connections in a multi-layer RNN, it is observed that the capability of the RNN to learn and model high-order dynamics and nonlinearities is significantly improved. The problem of vanishing/exploding gradient in space for a multilayer RNN is also alleviated using feedforward connections. These results are demonstrated using a quadrotor case study, for which a model of the altitude dynamics is learned with our particular network structure, while existing methods are unable to generalize as quickly or at all.

Nima Mohajerin, Steven L. Waslander

Dept. of Mechanical and Mechatronics Engineering University of Waterloo, Waterloo, ON,Canada

Recurrent Neural Networks Modular Neural Networks Robot Dynamics Learning

Aug 27, 2026

Rapid On-Robot Learning for Dynamic Manipulation Skills: Robot Juggling

We present an online learning framework that enables a bimanual robot to acquire diverse juggling patterns directly on physical hardware within minutes, even with a significant sim2real gap. One of the most important lessons from this work is that a model, even when far from reality, can be extremely useful for learning. This motivates a central philosophy of our approach: learning should build upon the robot's current knowledge rather than replace it. Our regularized memory-based learning puts this principle into practice by learning a local model from accumulated experience while retaining the global prior model to extrapolate where experience is sparse. This enables efficient and stable online learning from each new experience without resorting to uninformed exploration over a vast space of possible behaviors. Equally important to continual on-robot learning is safety, allowing the robot to repeatedly practice and improve in the real world. We construct a mutually reachable set that allows safe transitions between successive throws and catches, without driving either arm into a state from which its next action would require violating the robot's joint or actuator limits. Together, these ideas enable a bimanual robot with multi-fingered hands and onboard vision to safely learn and compose five canonical three-ball juggling patterns, including cascade, tennis, half-shower, shower, and box, within less than 5 minutes of real-world interaction. More broadly, this work points toward robots that build upon imperfect prior knowledge and continually refine their behavior through their own real-world experience.

Taeyoon Lee, Chunpeng Wang, Christopher G. Atkeson +2

Robotics and AI Institute (RAI), Cambridge MA, USA · Robotics Institute, Carnegie Mellon University, Pittsburgh PA, USA

Dynamic Robotic Manipulation Continual Robot Learning Robot Dynamics Learning Bimanual Robotic Manipulation

Aug 24, 2026

Reward-Free Continual Adaptation for Resilient Space Robots

Space robots operate in extreme environments where hardware degradation can critically compromise traditional control strategies. While continual reinforcement learning offers a promising mechanism for online adaptation, it inherently requires access to a reward signal during deployment. However, precise reward computation in space is often infeasible due to the lack of external tracking systems and the overall complexity of the environment. To address the challenge of unobservable rewards, we introduce a reward-free continual learning framework that leverages latent-state world models. By pre-training a model-based agent across diverse simulations, the world model learns a robust predictor of the reward structure within its latent space. Upon deployment to an environment with severe hardware degradation, we freeze the observation encoder and reward predictor to update only the transition dynamics of the world model through unsupervised rollouts. By training the policy entirely on imagined trajectories generated by this updated world model, the agent adapts to altered dynamics without receiving new rewards. We demonstrate our approach across simulated planetary traversal, orbital navigation, and precision assembly tasks subjected to severe morphological failures.

Andrej Orsula, Miguel Olivares-Mendez, Carol Martinez

University of Luxembourg

Reinforcement Learning Continual Robot Learning Reward-Free RL Space Robotics Model-Based RL Robot Dynamics Learning

Aug 21, 2026

NeSAM: Neuro-Symbolic Kinodynamics with Soil Adaptation for Off-Road Mobility

Accurate prediction of off-road vehicle motion over deformable terrain remains challenging because sinkage, slip, and traction vary with local soil conditions. Existing learning-based kinodynamic models directly approximate vehicle-terrain interactions from data but do not explicitly represent soil mechanics and offer limited physical interpretability. To address these limitations, we present NeSAM, a neuro-symbolic framework that combines differentiable Bekker-Wong terramechanics with learned terrain representations and a Transformer-based residual dynamics model for long-horizon, six degree-of-freedom kinodynamic prediction. The terramechanics component models soil-dependent interaction forces, while the residual model corrects discrepancies between the analytical prediction and the observed vehicle dynamics. NeSAM further estimates physically meaningful soil parameters from terrain observations and updates them online using an extended Kalman filter. We evaluate NeSAM in Verti-Bench, a simulator built on the Chrono multiphysics engine, and validate its performance on a physical Verti-4-Wheeler platform. NeSAM improves prediction accuracy by up to 30% in simulation and 29% on real-world data relative to the strongest compared baselines. When integrated with a close-loop navigation controller, NeSAM further improves traversal success rate through online soil adaptation while reduces Hausdorff distance to the reference trajectory by 69.4%, indicating improved trajectory tracking accuracy.

Chenhui Pan, Tong Xu, Francesco Cancelliere +1

Department of Computer Science, George Mason University, USA · Department of Electrical, Electronic and Computer Engineering, University of Catania, Italy

Terrain-Aware Robot Locomotion Robot Dynamics Learning Off-Road Robot Navigation Neurosymbolic Learning Residual Dynamics Learning

Aug 21, 2026

Graph-Operator World Models for Morphology-Parameter Generalization in Continuous Control

World models for continuous control are commonly trained for a fixed physical system and can degrade when known morphology parameters such as link lengths, masses, damping, and actuation change. Existing approaches often provide these parameters as conditioning information, but leave unspecified which part of the learned transition should remain reusable and which part should change with morphology. We propose Graph-Operator World Models (GraphOp-WM), a structured world model for generalization across unseen morphology parameters within related articulated robot families. GraphOp-WM represents bodies and their kinematic relations as an attributed graph and factorizes each transition into a morphology-independent local dynamics basis and a morphology-conditioned structured operator. The operator combines node-local modulation, kinematic-tree coupling, and a low-rank global correction, while architectural information separation, basis normalization, and paired-morphology supervision encourage static morphology dependence to be carried by the operator pathway. Graph-level readout and edge-wise action representations provide a compatible interface for reward, value, and TD-MPC-style planning. We further define controlled MuJoCo parameter splits covering interpolation, extrapolation, and held-out compositions of link geometry, mass, damping, and actuation parameters in Hopper, Walker2d, and HalfCheetah.

Xu Yang, Yiqin Yang, Qianchuan Zhao

Tsinghua University · The Key Laboratory of Cognition and Decision Intelligence for Complex Systems, Institute of Automation, Chinese Academy of Sciences

World Models for Robotics Model-Based RL Robot Dynamics Learning Action-Conditioned World Models

Aug 19, 2026

Learning the Right Abstraction: Neural Reduced Dynamics for Complex Robot Control

High-fidelity embodied AI simulators provide realistic evaluation of complex robotic systems, but their computational cost limits their direct use for large-scale reinforcement learning campaigns. We advocate the use of less accurate but more expeditious simulations, which might draw on data-driven, e.g., neural dynamics, models. This contribution argues that the practical value of a neural dynamics model for complex robot control lies in learning the right abstraction: a reduced state that preserves the control-relevant physics of the high-fidelity system while enabling high-throughput policy learning. We develop a neural reduced dynamics (NRD) framework that separates the state the model propagates from what can be supplied as an input or recovered analytically, trains policies entirely inside the frozen learned model, and validates them back in the high-fidelity simulator. Two case studies instantiate it across three control tasks: terrain-aware HMMWV trajectory tracking on rigid, bumpy and deformable Continuum Representation Model (CRM) terrain; and goal reaching for a stock tracked vehicle and its front-mounted articulated arm. Every policy transfers back to the high-fidelity simulator. A single policy trained inside the terrain-conditioned dynamics model, and given no terrain input of its own, attains lower median and mean tracking error than both single-terrain specialists on all three terrains, including zero-shot bumpy terrain. Quantitatively, the tracked vehicle reaches 100 of 100 goals and the arm 97 of 100, with zero contacts or joint-limit violations. The NRD models advance roughly four orders of magnitude faster in simulated time than the high-fidelity simulator scenes they replace, making iterative on-policy learning practical and supporting neural reduced dynamics as a bridge between accurate but expensive physics simulation and scalable robot learning.

Harry Zhang, Dan Negrut

Department of Mechanical Engineering, University of Wisconsin–Madison, Madison, 53706, WI, USA

Robotic Control Robot Dynamics Learning Robotics Simulation Reduced-Order Modeling Model-Based RL Neural Network Control

Aug 17, 2026

Orbit-Planner: Towards Latent World Models for On-Orbit Obstacle Avoidance of Satellite Agents

Satellite agents for on-orbit navigation tasks need to predict collision risks using limited onboard observations. However, conventional planners often rely on predefined maps and fixed environmental assumptions, limiting their adaptability in dynamic on-orbit scenarios. In this paper, we propose Orbit-Planner, a two-stage latent world model for on-orbit obstacle avoidance. Orbit-Planner learns action-conditioned spacecraft dynamics to perform future-state rollouts in latent space, and introduces a Physics Probe to decode physical state changes from imagined latent trajectories. Experiments demonstrate that Orbit-Planner can perform long-horizon latent rollouts and recover physical states from imagined trajectories. In closed-loop obstacle-avoidance navigation in Isaac Sim, it attains a success rate of 91.7%. Code is available at https://github.com/ZhijianLi2003/Orbit_Planner.

Zhijian Li, Chao Ren, Peijin Wang +1

Aerospace Information Research Institute, Chinese Academy of Sciences, 100094 Beijing, China · School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing 100190, China · University of Chinese Academy of Sciences, Beijing 100190, China +1

Collision Avoidance World Model-Based Planning Dynamic Obstacle Avoidance Robot Dynamics Learning

Jul 13, 2026

SKooP: Symmetric Koopman Predictions for Faster and More Generalizable Legged Robot Locomotion with Reinforcement Learning

Reinforcement learning (RL) algorithms classically suffer from poor sample efficiency. In robotics, a recent line of work has emerged addressing this problem by encoding physics priors in the learning process. However, most of these approaches are validated on well-defined, low-dimensional benchmark systems rather than high-dimensional robots with complex nonlinear dynamics. In this paper, we introduce SKooP (Symmetric Koopman Predictions), an approach combining the advantages of morphological symmetries with those of a Koopman model learned via autoencoder to enhance policy learning. SKooP learns a Koopman model of the system dynamics alongside the policy. The resulting Koopman predictions are used as privileged observations for the critic, allowing the agent to learn based on smoother, more informative features. We also incorporate group symmetries into the actor, critic, encoder and decoder networks to produce a highly equivariant policy. The SKooP approach is validated via in-depth analysis of the learned Koopman models and symmetric policies to showcase how each of these influences the agent's performance. We also show that the learned policies are transferable to different simulation environments. Our results show that SKooP consistently reduces convergence time and increases the learned reward for multiple challenging bipedal locomotion tasks on a quadruped robot. Project page: https://evelyd.github.io/SymmetricKoopmanPredictions

Evelyn D'Elia, Weishu Zhan, Giulio Turrisi +5

Machine Learning and Optimisation, University of Manchester, M13 9PL Manchester, U.K. · Dynamic Legged Systems Lab, IIT, 16163 Genoa, Italy · Generative Bionics S.R.L, 16163 Genoa, Italy +3

Reinforcement Learning Koopman Operator Learning Robot Dynamics Learning RL for Legged Locomotion Equivariant Policy Learning Robot Policy Generalization

Jun 30, 2026

3D Point World Models: Point Completion Enables More Accurate Dynamics Learning

Learning predictive models of the world enables robotic control through planning, potentially allowing robots to improvise solutions on new tasks. However, large video-based dynamics models lack explicit 3D spatial structure and suffer from geometrically inconsistent long-term rollouts with compounding errors. Emerging 3D dynamics models based on partial point clouds improve geometric consistency but remain sensitive to occlusions and accumulated prediction drift. To address these challenges, we present 3D Point World Models (3DPWM) - a task-agnostic world model that operates entirely in 3D space by first completing partial point clouds and then learning action-conditioned dynamics in this completed 3D scene. By operating on completed geometry, 3DPWM enables reliable long-horizon rollouts and more accurate cost evaluation for model-based planning while supporting adaptation to new tasks. Experiments across different robotic embodiments and tabletop manipulation benchmarks demonstrate that 3DPWM achieves significantly more reliable long-horizon rollouts (100-300+ steps), supports both open-loop and closed-loop planning, and enables successful sim-to-real transfer.

Skand Peri, Hung Nguyen, Chanho Kim +2

Oregon State University

World Model Learning World Models for Robotics World Model-Based Planning Robot Dynamics Learning Action-Conditioned World Models Robotic Manipulation+1

Jun 25, 2026

Continual Robot Policy Learning via Variational Neural Dynamics

Robots deployed in the real world rarely operate under a single fixed dynamics model: wind changes, payloads vary, batteries drain, contacts shift, and hardware wears. Yet most learning-based controllers are trained once and deployed as if learning were complete. This prevents the robot from using deployment experience to further improve task performance. In this work, we propose a continual learning framework that uses real-world experience to improve robot policies under hidden and recurring dynamics. Our method learns a condition-aware dynamics model from real state-action trajectories by combining an analytical physics prior with a neural residual for unmodeled effects. A recurrent encoder infers the current hidden condition from recent interaction, and this estimate conditions both the residual model and the policy. Policy learning is performed via differentiable simulation using diverse learned dynamics sampled from the latent model. At deployment, these sampled conditions are replaced by conditions inferred online from recent real interaction, allowing the policy to recover recurring dynamics by recognition rather than residual re-fitting. Through extensive simulation studies and real-world experiments, we demonstrate that the framework improves policy performance under diverse unobserved disturbances. On real quadrotor trajectory tracking under changing wind, the policy recovers from recurring disturbances in roughly 1s, about 5x faster than online residual re-fitting. It also reduces large-disturbance hover and tracking errors by 65.7% and 53.3% over the state-of-the-art online adaptation approaches

Jiaxu Xing, Zhiyuan Zhu, Yunfan Ren +4

Robotics and Perception Group, University of Zurich

Robot Policy Learning Robot Policy Adaptation Continual Robot Learning Residual Dynamics Learning Robot Dynamics Learning Robust RL+3

Jun 15, 2026

Unified Motion-Action Modeling for Heterogeneous Robot Learning

We present Unified Motion-Action (UMA) Model, an approach that uses 3D object motion trajectories as a shared interface to bridge visuomotor control and dynamics modeling. UMA treats object motion and robot actions as co-evolving variables under a masked generative objective, in which the mask pattern determines both the supervision regime during pretraining and the inference mode at deployment. Using hindsight-relabeled motion contexts and a contrastive objective that disentangles task intent from scene geometry, UMA enables multi-task pretraining across heterogeneous data sources without requiring manually annotated task instructions. At deployment, the same pretrained parameters support motion-conditioned visuomotor control, motion-based dynamics modeling, and task adaptation from few-shot demonstrations. Pretrained on a mixture of robot demonstrations, human videos, and simulated data, UMA consistently outperforms state-of-the-art baselines specialized for each inference mode.

Yunhao Cao, Shitong Liu, Chao Feng +4

Cornell University

Robotic Control Robot Dynamics Learning Visuomotor Policy Learning Cross-Embodiment Robot Learning Conditional Generative Modeling Robot Learning

Jun 13, 2026

Learning Context-Aware Neural ODE Dynamics for Adaptive Robotic Control

Robotic systems deployed in uncertain and dynamically changing environments often face variations in contact conditions, aerodynamic effects, and external disturbances that challenge reliable control. To remain effective under model-based control, these systems require dynamics models that can adapt to such changes, especially when direct access to complete environmental information is limited. To enable adaptability and facilitate integration with model predictive control, we propose a context-aware dynamics model based on neural ordinary differential equations, which infers environmental factors from state-action histories using a two-phase training procedure. We validate the approach across diverse robotic platforms, including a quadrotor in simulation, as well as a Sphero BOLT robot and a Fanuc manipulator in real-world experiments. The results demonstrate that our method effectively adapts to temporally and spatially varying environmental changes across different tasks. Videos are available at https://youtu.be/PY0sNyF2rqE , and the source code is available at https://github.com/syyu410-yu/context-aware-neural-ode-control.git .

Shao-Yi Yu, Jen-Wei Wang, Maya Horii +2

Department of Mechanical Engineering, University of California, Berkeley, CA, USA · Aalto University and YaiYai Ltd, Finland

Robot Dynamics Identification Robot Dynamics Learning Robotics Neural ODEs Adaptive Control

Jun 8, 2026

Physics-Aware Sparse Learning and Selective Online Adaptation for Euler-Lagrange Robot Dynamics

Accurate dynamics models are essential for model-based robotic control, yet nominal Euler--Lagrange models often become inaccurate in the presence of payload variation, unmodeled coupling, friction, aerodynamic effects, and changing operating conditions. Most learning-based correction methods improve prediction accuracy by introducing a single additive residual, but do not preserve the internal mechanical structure of Euler--Lagrange systems. This leads to models that do not preserve symmetry, positive-definiteness, or the coupling between inertia and velocity-dependent terms, which can result in physically inconsistent predictions and reduced reliability when embedded in model-based controllers. We propose a structure-preserving residual learning framework that decomposes model mismatch into an inertia correction, the corresponding induced Coriolis term, and a generalized-force residual. The mechanical component is learned under physical constraints, while the disturbance-sensitive component is represented through a sparse history-dependent latent interaction model and adapted online using Bayesian linear regression. This separation preserves key mechanical structure while restricting adaptation to the part of the dynamics most affected by changing conditions. Experiments across multiple robotic platforms, including mobile, aerial, and manipulator systems, show that the proposed method improves dynamics prediction and trajectory tracking under coupled and time-varying dynamics. These results highlight the value of combining structured residual modeling, compact latent interaction selection, and selective online adaptation for real-world model-based control.

Rishabh Dev Yadav, Samaksh Ujjawal, Sihao Sun +2

Department of Computer Science, The University of Manchaster, U.K. · Department of Cognitive Robotics, Delft University of Technology, Netherlands · International Institute of Information Technology Hyderabad. +1

Robotic Control Robot Dynamics Identification Online System Identification Residual Learning Residual Dynamics Learning Robot Dynamics Learning+2

Jun 1, 2026

Dynamics Are Learned, Not Told: Semi-Supervised Discovery of Latent Dynamics Geometries For Zero-Shot Policy Adaptation

Real-world dynamics shifts pose a critical challenge for reinforcement learning in robotics, as policies tightly coupled to nominal environments often fail catastrophically when physical conditions change. Most existing methods rely on encoding explicitly identified physical parameters into a latent context, a parameter-centric paradigm that depends on pre-specified axes of variation and becomes brittle under unmodeled or compound dynamics changes. We revisit dynamics adaptation from an outcome-centric perspective: rather than telling policies what the dynamics are, we enable them to learn how dynamics affect interaction outcomes. Theoretically, this is grounded in a monotonic relationship between target-domain regret and the Lipschitz constant of a trajectory dynamics encoder. Practically, this constant can be upper-bounded through contrastive learning, yielding a smooth, task-relevant latent topology without privileged dynamics information. On MuJoCo benchmarks, our method consistently outperforms parameter-centric baselines under severe dynamics shifts, including unmodeled and time-varying parameters, while also improving in-distribution stability and latent interpretability. Overall, these results validate that controlling latent geometry is a principled mechanism for robust adaptation.

Zhiming Xu, Weitao Zhou, Xianghui Pan +4

School of Computer Science and Engineering, Tongji University, Shanghai, China · School of Vehicle and Mobility, Tsinghua University, Beijing, China · Simple AI, Beijing, China +1

Robot Policy Adaptation Latent Dynamics Modeling Robot Dynamics Learning Robotic RL Zero-Shot Policy Transfer Robust RL

May 29, 2026

Batched Differentiable Rigid Body Dynamics in PyTorch for GPU-Accelerated Robot Learning

As robot control shifts toward large-scale reinforcement learning with in-loop dynamics computation, the community's reliance on CPU-bound libraries such as Pinocchio creates a throughput bottleneck in GPU-based training pipelines. We present BARD (Batched Articulated Rigid-body Dynamics), a self-contained PyTorch implementation of Featherstone's rigid-body dynamics algorithms, optimized for batched GPU evaluation and automatic differentiation. Three design choices make this efficient: a tiered lazy-evaluation cache that avoids redundant tree traversals, matmul-free joint transforms via pre-computed Rodrigues constants, and level-parallel propagation that reduces sequential operations to tree-depth batched steps. On five robot models (7-23 DOFs), BARD matches Pinocchio numerically while reaching up to 64x higher throughput for Forward Kinematics and 63x for Jacobians at batch size 4096 on an NVIDIA H200. We validate differentiability through gradient-based system identification on a 7-DOF manipulator, recovering link masses to 1.24% mean error under 5% torque noise, and integrate BARD into an Isaac Lab AMP training pipeline for an 11-DOF spined quadruped with 4096 parallel environments, where it is 8.5x faster than Pinocchio and 2.0x faster than ADAM for in-loop dynamics. BARD is open-sourced at: https://github.com/YueWang996/bard-pytorch-dynamics.

Yue Wang, Yanran Xu, Wenbo Wu +2

University of Southampton, Southampton, SO17 1BJ, UK

Multibody Dynamics Robot Dynamics Learning Differentiable Programming Differentiable Physics Robot Learning

May 2, 2026

Dynamics Aware Quadrupedal Locomotion via Intrinsic Dynamics Head

Quadrupedal locomotion plays a critical role in enabling agile, versatile movement across complex terrains. Understanding and estimating the underlying physical dynamics are essential for achieving efficient and stable quadrupedal locomotion. We propose a novel training framework for quadrupedal locomotion that enables the Control Policy to understand and reason about physical dynamics. In simulation, we concurrently train an Intrinsic Dynamics (ID) Head that learns state-to-torque dynamics alongside the Control Policy, and we define a dynamics reward enabled by the ID Head that encourages the Policy toward more predictable dynamical behavior. We also provide a mechanism to tune the learned dynamics in the resulting Policy by controlling the training coefficients of the ID Head. Our simulation experiments show that this mechanism drives convergence to better optima across a wide range of standard quadrupedal locomotion rewards, yielding more efficient and smoother policies. Our real-robot experiments demonstrate sim-to-real transfer of these improvements, with significant gains in torque efficiency (16.8%), action rate (18.6%), and mechanical power (12.8%), while improving safe torque occupancy by 6.4%.

Aman Arora, Nalini Ratha

Department of Computer Science, University at Buffalo, The State University of New York, Buffalo, USA

Robot Dynamics Learning RL for Legged Locomotion Inverse Dynamics Quadruped Locomotion Robot Learning

Apr 28, 2026

Learning-Based Dynamics Modeling and Robust Control for Tendon-Driven Continuum Robots

Tendon-Driven Continuum Robots (TDCRs) pose significant modeling and control challenges due to complex nonlinearities, such as frictional hysteresis and transmission compliance. This paper proposes a differentiable learning framework that integrates high-fidelity dynamics modeling with robust neural control. We develop a GRU-based dynamics model featuring bidirectional multi-channel connectivity and residual prediction to effectively suppress compounding errors during long-horizon auto-regressive prediction. By treating this model as a gradient bridge, an end-to-end neural control policy is optimized through backpropagation, allowing it to implicitly internalize compensation for intricate nonlinearities. Experimental validation on a physical three-section TDCR demonstrates that our framework achieves accurate tracking and superior robustness against unseen payloads, outperforming Jacobian-based methods by eliminating self-excited oscillations. For implementation details and source code, please refer to https://github.com/ZiqingZou/ContinuumControl.

Ziqing Zou, Ke Qiu, Fei Wang +3

Department of Control Science and Engineering, Zhejiang University, Hangzhou, China

Gated Recurrent Unit Robot Trajectory Tracking Robot Dynamics Learning Recurrent Neural Networks Robust Control Continuum Robot Control+2

Apr 21, 2026

Efficient Reinforcement Learning using Linear Koopman Dynamics for Nonlinear Robotic Systems

This paper presents a model-based reinforcement learning (RL) framework for optimal closed-loop control of nonlinear robotic systems. The proposed approach learns linear lifted dynamics through Koopman operator theory and integrates the resulting model into an actor-critic architecture for policy optimization, where the policy represents a parameterized closed-loop controller. To reduce computational cost and mitigate model rollout errors, policy gradients are estimated using one-step predictions of the learned dynamics rather than multi-step propagation. This leads to an online mini-batch policy gradient framework that enables policy improvement from streamed interaction data. The proposed framework is evaluated on several simulated nonlinear control benchmarks and two real-world hardware platforms, including a Kinova Gen3 robotic arm and a Unitree Go1 quadruped. Experimental results demonstrate improved sample efficiency over model-free RL baselines, superior control performance relative to model-based RL baselines, and control performance comparable to classical model-based methods that rely on exact system dynamics.

Wenjian Hao, Yuxuan Fang, Zehui Lu +1

School of Aeronautics and Astronautics, Purdue University, West Lafayette, IN 47907, USA · independent researcher

Robotic Control Reinforcement Learning Koopman Operator Learning Robot Dynamics Learning Actor-Critic Methods Optimal Control+1

Apr 15, 2026

Diffusion Sequence Models for Generative In-Context Meta-Learning of Robot Dynamics

Accurate modeling of robot dynamics is essential for model-based control, yet remains challenging under distributional shifts and real-time constraints. In this work, we formulate system identification as an in-context meta-learning problem and compare deterministic and generative sequence models for forward dynamics prediction. We take a Transformer-based meta-model, as a strong deterministic baseline, and introduce to this setting two complementary diffusion-based approaches: (i) inpainting diffusion (Diffuser), which learns the joint input-observation distribution, and (ii) conditioned diffusion models (CNN and Transformer), which generate future observations conditioned on control inputs. Through large-scale randomized simulations, we analyze performance across in-distribution and out-of-distribution regimes, as well as computational trade-offs relevant for control. We show that diffusion models significantly improve robustness under distribution shift, with inpainting diffusion achieving the best performance in our experiments. Finally, we demonstrate that warm-started sampling enables diffusion models to operate within real-time constraints, making them viable for control applications. These results highlight generative meta-models as a promising direction for robust system identification in robotics.

Angelo Moroncelli, Matteo Rufolo, Gunes Cagin Aydin +2

University of Applied Science and Arts of Southern Switzerland, Department of Innovative Technologies, IDSIA-SUPSI, Lugano, Switzerland. · Università della Svizzera Italiana, Faculty of Informatics, Lugano, Switzerland. · Politecnico di Milano, Mechanical Department, Milano, Italy.

Robot Dynamics Identification Robot System Identification Robot Dynamics Learning System Identification Diffusion Models ICL for Robotics

Apr 10, 2026

Physics-Informed Reinforcement Learning of Spatial Density Velocity Potentials for Map-Free Racing

Autonomous racing without prebuilt maps is a grand challenge for embedded robotics that requires kinodynamic planning from instantaneous sensor data at the acceleration and tire friction limits. Out-Of-Distribution (OOD) generalization to various racetrack configurations utilizes Machine Learning (ML) to encode the mathematical relation between sensor data and vehicle actuation for end-to-end control, with implicit localization. These comprise Behavioral Cloning (BC) that is capped to human reaction times and Deep Reinforcement Learning (DRL) which requires large-scale collisions for comprehensive training that can be infeasible without simulation but is arduous to transfer to reality, thus exhibiting greater performance than BC in simulation, but actuation instability on hardware. This paper presents a DRL method that parameterizes nonlinear vehicle dynamics from the spectral distribution of depth measurements with a non-geometric, physics-informed reward, to infer vehicle time-optimal and overtaking racing controls with an Artificial Neural Network (ANN) that utilizes less than 1% of the computation of BC and model-based DRL. Slaloming from simulation to reality transfer and variance-induced conservatism are eliminated with the combination of a physics engine exploit-aware reward and the replacement of an explicit collision penalty with an implicit truncation of the value horizon. The policy outperforms human demonstrations by 12% in OOD tracks on proportionally scaled hardware, by maximizing the friction circle with tire dynamics that resemble an empirical Pacejka tire model. System identification illuminates a functional bifurcation where the first layer compresses spatial observations to extract digitized track features with higher resolution in corner apexes, and the second encodes nonlinear dynamics.

Shathushan Sivashangaran, Apoorva Khairnar, Sepideh Gohari +2

Autonomous Robots and Vehicles Laboratory, College of Engineering, Virginia Commonwealth University, Richmond, VA 23284, USA. · Department of Mechanical Engineering, Virginia Tech, Blacksburg, VA 24061, USA. · Robotics Department, University of Michigan, Ann Arbor, MI 48109, USA.

Autonomous Racing RL Control Physics-Informed RL Robot Dynamics Learning

Apr 9, 2026

Morphology-Conditioned World Model for Cross-Embodiment Quadrupedal Locomotion

World models promise a paradigm shift in robotics, where an agent learns the physics of its environment once and then acquires behaviors efficiently. Yet the learned dynamics models at their core are typically morphology locked. In legged locomotion, a dynamics model trained on an ANYmal-D quadruped fails on a Unitree Go1 because it overfits to one robot's embodiment rather than capturing the locomotion dynamics shared across robots, so even a small change in actuator dynamics or limb length forces retraining from scratch. However, if we formalize a robot's unique physical traits into a morphology specification, a controller for a family of robots can utilize this blueprint in two ways. It can feed the specification to a model-free policy, or it can feed the specification to a learned dynamics model and extract the policy in imagination. We argue for the second route and introduce the Quadrupedal World Model (QWM), which conditions a single generative dynamics model on scale-invariant physical features and trains policies entirely inside it, through a physical morphology encoder, an adaptive reward normalizer, and morphology conditioning in the latent dynamics. Holding the morphology information identical, a model-free policy matches QWM on the training cohort but degrades on unseen morphologies, while QWM transfers zero-shot with no fine-tuning, adaptation, or warm-up in such cases. To our knowledge, this is the first world model to demonstrate zero-shot cross-embodiment transfer within the quadrupedal family.

Mohamad H. Danesh, Chenhao Li, Amin Abyaneh +5

McGill University, Canada · Mila - Quebec AI Institute, Canada · ETH Zurich, Switzerland +1

World Models for Robotics Cross-Embodiment Robot Learning World Model Learning Cross-Embodiment Transfer Robot Dynamics Learning Quadruped Locomotion