Marco Hutter

Mayank Mittal, Calvin Yu, Qinxi Yu et al. · cmu, eth-zurich

We present Orbit, a unified and modular framework for robot learning powered by NVIDIA Isaac Sim. It offers a modular design to easily and efficiently create robotic environments with photo-realistic scenes and high-fidelity rigid and deformable body simulation. With Orbit, we provide a suite of benchmark tasks of varying difficulty -- from single-stage cabinet opening and cloth folding to multi-stage tasks such as room reorganization. To support working with diverse observations and action spaces, we include fixed-arm and mobile manipulators with different physically-based sensors and motion generators. Orbit allows training reinforcement learning policies and collecting large demonstration datasets from hand-crafted or expert solutions in a matter of minutes by leveraging GPU-based parallelization. In summary, we offer an open-sourced framework that readily comes with 16 robotic platforms, 4 sensor modalities, 10 motion generators, more than 20 benchmark tasks, and wrappers to 4 learning libraries. With this framework, we aim to support various research areas, including representation learning, reinforcement learning, imitation learning, and task and motion planning. We hope it helps establish interdisciplinary collaborations in these communities, and its modularity makes it easily extensible for more tasks and applications in the future.

Mayank Mittal, Pascal Roth, James Tigue et al. · nvidia

We present Isaac Lab, the natural successor to Isaac Gym, which extends the paradigm of GPU-native robotics simulation into the era of large-scale multi-modal learning. Isaac Lab combines high-fidelity GPU parallel physics, photorealistic rendering, and a modular, composable architecture for designing environments and training robot policies. Beyond physics and rendering, the framework integrates actuator models, multi-frequency sensor simulation, data collection pipelines, and domain randomization tools, unifying best practices for reinforcement and imitation learning at scale within a single extensible platform. We highlight its application to a diverse set of challenges, including whole-body control, cross-embodiment mobility, contact-rich and dexterous manipulation, and the integration of human demonstrations for skill acquisition. Finally, we discuss upcoming integration with the differentiable, GPU-accelerated Newton physics engine, which promises new opportunities for scalable, data-efficient, and gradient-based approaches to robot learning. We believe Isaac Lab's combination of advanced simulation capabilities, rich sensing, and data-center scale execution will help unlock the next generation of breakthroughs in robotics research.

Nikita Rudin, David Hoeller, Marko Bjelonic et al. · eth-zurich

The common approach for local navigation on challenging environments with legged robots requires path planning, path following and locomotion, which usually requires a locomotion control policy that accurately tracks a commanded velocity. However, by breaking down the navigation problem into these sub-tasks, we limit the robot's capabilities since the individual tasks do not consider the full solution space. In this work, we propose to solve the complete problem by training an end-to-end policy with deep reinforcement learning. Instead of continuously tracking a precomputed path, the robot needs to reach a target position within a provided time. The task's success is only evaluated at the end of an episode, meaning that the policy does not need to reach the target as fast as possible. It is free to select its path and the locomotion gait. Training a policy in this way opens up a larger set of possible solutions, which allows the robot to learn more complex behaviors. We compare our approach to velocity tracking and additionally show that the time dependence of the task reward is critical to successfully learn these new behaviors. Finally, we demonstrate the successful deployment of policies on a real quadrupedal robot. The robot is able to cross challenging terrains, which were not possible previously, while using a more energy-efficient gait and achieving a higher success rate.

Eric Vollenweider, Marko Bjelonic, Victor Klemm et al. · eth-zurich

In recent years, reinforcement learning (RL) has shown outstanding performance for locomotion control of highly articulated robotic systems. Such approaches typically involve tedious reward function tuning to achieve the desired motion style. Imitation learning approaches such as adversarial motion priors aim to reduce this problem by encouraging a pre-defined motion style. In this work, we present an approach to augment the concept of adversarial motion prior-based RL to allow for multiple, discretely switchable styles. We show that multiple styles and skills can be learned simultaneously without notable performance differences, even in combination with motion data-free skills. Our approach is validated in several real-world experiments with a wheeled-legged quadruped robot showing skills learned from existing RL controllers and trajectory optimization, such as ducking and walking, and novel skills such as switching between a quadrupedal and humanoid configuration. For the latter skill, the robot is required to stand up, navigate on two wheels, and sit down. Instead of tuning the sit-down motion, we verify that a reverse playback of the stand-up movement helps the robot discover feasible sit-down behaviors and avoids tedious reward function tuning.

Robin Schmid, Deegan Atha, Frederik Schöller et al. · eth-zurich

Navigating off-road with a fast autonomous vehicle depends on a robust perception system that differentiates traversable from non-traversable terrain. Typically, this depends on a semantic understanding which is based on supervised learning from images annotated by a human expert. This requires a significant investment in human time, assumes correct expert classification, and small details can lead to misclassification. To address these challenges, we propose a method for predicting high- and low-risk terrains from only past vehicle experience in a self-supervised fashion. First, we develop a tool that projects the vehicle trajectory into the front camera image. Second, occlusions in the 3D representation of the terrain are filtered out. Third, an autoencoder trained on masked vehicle trajectory regions identifies low- and high-risk terrains based on the reconstruction error. We evaluated our approach with two models and different bottleneck sizes with two different training and testing sites with a fourwheeled off-road vehicle. Comparison with two independent test sets of semantic labels from similar terrain as training sites demonstrates the ability to separate the ground as low-risk and the vegetation as high-risk with 81.1% and 85.1% accuracy.

Jesus Tordesillas, Jonathan P. How, Marco Hutter · eth-zurich

This paper presents RAYEN, a framework to impose hard convex constraints on the output or latent variable of a neural network. RAYEN guarantees that, for any input or any weights of the network, the constraints are satisfied at all times. Compared to other approaches, RAYEN does not perform a computationally-expensive orthogonal projection step onto the feasible set, does not rely on soft constraints (which do not guarantee the satisfaction of the constraints at test time), does not use conservative approximations of the feasible set, and does not perform a potentially slow inner gradient descent correction to enforce the constraints. RAYEN supports any combination of linear, convex quadratic, second-order cone (SOC), and linear matrix inequality (LMI) constraints, achieving a very small computational overhead compared to unconstrained networks. For example, it is able to impose 1K quadratic constraints on a 1K-dimensional variable with an overhead of less than 8 ms, and an LMI constraint with 300x300 dense matrices on a 10K-dimensional variable in less than 12 ms. When used in neural networks that approximate the solution of constrained optimization problems, RAYEN achieves computation times between 20 and 7468 times faster than state-of-the-art algorithms, while guaranteeing the satisfaction of the constraints at all times and obtaining a cost very close to the optimal one.

David Hoeller, Nikita Rudin, Christopher Choy et al. · eth-zurich

We propose a learning-based method to reconstruct the local terrain for locomotion with a mobile robot traversing urban environments. Using a stream of depth measurements from the onboard cameras and the robot's trajectory, the algorithm estimates the topography in the robot's vicinity. The raw measurements from these cameras are noisy and only provide partial and occluded observations that in many cases do not show the terrain the robot stands on. Therefore, we propose a 3D reconstruction model that faithfully reconstructs the scene, despite the noisy measurements and large amounts of missing data coming from the blind spots of the camera arrangement. The model consists of a 4D fully convolutional network on point clouds that learns the geometric priors to complete the scene from the context and an auto-regressive feedback to leverage spatio-temporal consistency and use evidence from the past. The network can be solely trained with synthetic data, and due to extensive augmentation, it is robust in the real world, as shown in the validation on a quadrupedal robot, ANYmal, traversing challenging settings. We run the pipeline on the robot's onboard low-power computer using an efficient sparse tensor implementation and show that the proposed method outperforms classical map representations.

Álvaro Belmonte-Baeza, Joonho Lee, Giorgio Valsecchi et al. · eth-zurich

The process of robot design is a complex task and the majority of design decisions are still based on human intuition or tedious manual tuning. A more informed way of facing this task is computational design methods where design parameters are concurrently optimized with corresponding controllers. Existing approaches, however, are strongly influenced by predefined control rules or motion templates and cannot provide end-to-end solutions. In this paper, we present a design optimization framework using model-free meta reinforcement learning, and its application to the optimizing kinematics and actuator parameters of quadrupedal robots. We use meta reinforcement learning to train a locomotion policy that can quickly adapt to different designs. This policy is used to evaluate each design instance during the design optimization. We demonstrate that the policy can control robots of different designs to track random velocity commands over various rough terrains. With controlled experiments, we show that the meta policy achieves close-to-optimal performance for each design instance after adaptation. Lastly, we compare our results against a model-based baseline and show that our approach allows higher performance while not being constrained by predefined motions or gait patterns.

Julian Nubert, Etienne Walther, Shehryar Khattak et al. · eth-zurich

LiDAR-based localization and mapping is one of the core components in many modern robotic systems due to the direct integration of range and geometry, allowing for precise motion estimation and generation of high quality maps in real-time. Yet, as a consequence of insufficient environmental constraints present in the scene, this dependence on geometry can result in localization failure, happening in self-symmetric surroundings such as tunnels. This work addresses precisely this issue by proposing a neural network-based estimation approach for detecting (non-)localizability during robot operation. Special attention is given to the localizability of scan-to-scan registration, as it is a crucial component in many LiDAR odometry estimation pipelines. In contrast to previous, mostly traditional detection approaches, the proposed method enables early detection of failure by estimating the localizability on raw sensor measurements without evaluating the underlying registration optimization. Moreover, previous approaches remain limited in their ability to generalize across environments and sensor types, as heuristic-tuning of degeneracy detection thresholds is required. The proposed approach avoids this problem by learning from a collection of different environments, allowing the network to function over various scenarios. Furthermore, the network is trained exclusively on simulated data, avoiding arduous data collection in challenging and degenerate, often hard-to-access, environments. The presented method is tested during field experiments conducted across challenging environments and on two different sensor types without any modifications. The observed detection performance is on par with state-of-the-art methods after environment-specific threshold tuning.

Benjamin Bokser, Daniel Gonzalez, Aaron Preston et al. · mit

Trials cyclists and mountain bike riders can hop, jump, balance, and drive on one or both wheels. This versatility allows them to achieve speed and energy-efficiency on smooth terrain and agility over rough terrain. Inspired by these athletes, we present the design and control of a robotic platform, Ultra Mobility Vehicle (UMV), which combines a bicycle and a reaction mass to move dynamically with minimal actuated degrees of freedom. We employ a simulation-driven design optimization process to synthesize a spatial linkage topology with a focus on vertical jump height and momentum-based balancing on a single wheel contact. Using a constrained Reinforcement Learning (RL) framework, we demonstrate zero-shot transfer of diverse athletic behaviors, including track-stands, jumps, wheelies, rear wheel hopping, and front flips. This 23.5 kg robot is capable of high speeds (8 m/s) and jumping on and over large obstacles (1 m tall, or 130% of the robot's nominal height).

Manthan Patel, Gabriel Waibel, Shehryar Khattak et al. · eth-zurich

Detecting objects of interest, such as human survivors, safety equipment, and structure access points, is critical to any search-and-rescue operation. Robots deployed for such time-sensitive efforts rely on their onboard sensors to perform their designated tasks. However, as disaster response operations are predominantly conducted under perceptually degraded conditions, commonly utilized sensors such as visual cameras and LiDARs suffer in terms of performance degradation. In response, this work presents a method that utilizes the complementary nature of vision and depth sensors to leverage multi-modal information to aid object detection at longer distances. In particular, depth and intensity values from sparse LiDAR returns are used to generate proposals for objects present in the environment. These proposals are then utilized by a Pan-Tilt-Zoom (PTZ) camera system to perform a directed search by adjusting its pose and zoom level for performing object detection and classification in difficult environments. The proposed work has been thoroughly verified using an ANYmal quadruped robot in underground settings and on datasets collected during the DARPA Subterranean Challenge finals.

Marco Camurri, Enrico Tomelleri, Matías Mattamala et al. · oxford

Covering one third of Earth's land surface, forests are vital to global biodiversity, climate regulation, and human well-being. In Europe, forests and woodlands reach approximately 40% of land area, and the forestry sector is central to achieving the EU's climate neutrality and biodiversity goals; these emphasize sustainable forest management, increased use of long-lived wood products, and resilient forest ecosystems. To meet these goals and properly address their inherent challenges, current practices require further innovation. This chapter introduces DigiForest, a novel, large-scale precision forestry approach leveraging digital technologies and autonomous robotics. DigiForest is structured around four main components: (1) autonomous, heterogeneous mobile robots (aerial, legged, and marsupial) for tree-level data collection; (2) automated extraction of tree traits to build forest inventories; (3) a Decision Support System (DSS) for forecasting forest growth and supporting decision-making; and (4) low-impact selective logging using purpose-built autonomous harvesters. These technologies have been extensively validated in real-world conditions in several locations, including forests in Finland, the UK, and Switzerland.

Julia Richter, David Oberacker, Gabriela Ligeza et al.

Robotic prospecting for critical resources on the Moon, such as ilmenite, rare earth elements, and water ice, requires robust exploration methods given the diverse terrain and harsh environmental conditions. Although numerous analog field trials address these goals, comparing their results remains challenging because of differences in robot platforms and experimental setups. These missions typically assess performance using selected, scenario-specific engineering metrics that fail to establish a clear link between field performance and science-driven objectives. In this paper, we address this gap by deriving a structured framework of KPI from three realistic multi-robot lunar scenarios reflecting scientific objectives and operational constraints. Our framework emphasizes scenario-dependent priorities in efficiency, robustness, and precision, and is explicitly designed for practical applicability in field deployments. We validated the framework in a multi-robot field test and found it practical and easy to apply for efficiency- and robustness-related KPI, whereas precision-oriented KPI require reliable ground-truth data that is not always feasible to obtain in outdoor analog environments. Overall, we propose this framework as a common evaluation standard enabling consistent, goal-oriented comparison of multi-robot field trials and supporting systematic development of robotic systems for future planetary exploration.

Fabian Jenelten, Junzhe He, Farbod Farshidian et al.

Legged locomotion is a complex control problem that requires both accuracy and robustness to cope with real-world challenges. Legged systems have traditionally been controlled using trajectory optimization with inverse dynamics. Such hierarchical model-based methods are appealing due to intuitive cost function tuning, accurate planning, generalization, and most importantly, the insightful understanding gained from more than one decade of extensive research. However, model mismatch and violation of assumptions are common sources of faulty operation. Simulation-based reinforcement learning, on the other hand, results in locomotion policies with unprecedented robustness and recovery skills. Yet, all learning algorithms struggle with sparse rewards emerging from environments where valid footholds are rare, such as gaps or stepping stones. In this work, we propose a hybrid control architecture that combines the advantages of both worlds to simultaneously achieve greater robustness, foot-placement accuracy, and terrain generalization. Our approach utilizes a model-based planner to roll out a reference motion during training. A deep neural network policy is trained in simulation, aiming to track the optimized footholds. We evaluate the accuracy of our locomotion pipeline on sparse terrains, where pure data-driven methods are prone to fail. Furthermore, we demonstrate superior robustness in the presence of slippery or deformable ground when compared to model-based counterparts. Finally, we show that our proposed tracking controller generalizes across different trajectory optimization methods not seen during training. In conclusion, our work unites the predictive capabilities and optimality guarantees of online planning with the inherent robustness attributed to offline learning.

Arjun Bhardwaj, Jonas Rothfuss, Bhavya Sukhija et al.

We introduce PACOH-RL, a novel model-based Meta-Reinforcement Learning (Meta-RL) algorithm designed to efficiently adapt control policies to changing dynamics. PACOH-RL meta-learns priors for the dynamics model, allowing swift adaptation to new dynamics with minimal interaction data. Existing Meta-RL methods require abundant meta-learning data, limiting their applicability in settings such as robotics, where data is costly to obtain. To address this, PACOH-RL incorporates regularization and epistemic uncertainty quantification in both the meta-learning and task adaptation stages. When facing new dynamics, we use these uncertainty estimates to effectively guide exploration and data collection. Overall, this enables positive transfer, even when access to data from prior tasks or dynamic settings is severely limited. Our experiment results demonstrate that PACOH-RL outperforms model-based RL and model-based Meta-RL baselines in adapting to new dynamic conditions. Finally, on a real robotic car, we showcase the potential for efficient RL policy adaptation in diverse, data-scarce conditions.

Fang Nan, Hao Ma, Qinghua Guan et al.

We present an online model-based reinforcement learning algorithm suitable for controlling complex robotic systems directly in the real world. Unlike prevailing sim-to-real pipelines that rely on extensive offline simulation and model-free policy optimization, our method builds a dynamics model from real-time interaction data and performs policy updates guided by the learned dynamics model. This efficient model-based reinforcement learning scheme significantly reduces the number of samples to train control policies, enabling direct training on real-world rollout data. This significantly reduces the influence of bias in the simulated data, and facilitates the search for high-performance control policies. We adopt online optimization analysis to derive sublinear regret bounds under stochastic online optimization assumptions, providing formal guarantees on performance improvement as more interaction data are collected. Experimental evaluations were performed on a hydraulic excavator arm and a soft robot arm, where the algorithm demonstrates strong sample efficiency compared to model-free reinforcement learning methods, reaching comparable performance within hours. Robust adaptation to shifting dynamics was also observed when the payload condition was randomized. Our approach paves the way toward efficient and reliable on-robot learning for a broad class of challenging control tasks.

Tao Zhang, Kaixian Qu, Zhibin Li et al.

Large language models are increasingly used as planners for robotic systems, yet how safely they plan remains an open question. To evaluate safe planning systematically, we introduce DESPITE, a benchmark of 12,279 tasks spanning physical and normative dangers with fully deterministic validation. Across 23 models, even near-perfect planning ability does not ensure safety: the best-planning model fails to produce a valid plan on only 0.4% of tasks but produces dangerous plans on 28.3%. Among 18 open-source models from 3B to 671B parameters, planning ability improves substantially with scale (0.4-99.3%) while safety awareness remains relatively flat (38-57%). We identify a multiplicative relationship between these two capacities, showing that larger models complete more tasks safely primarily through improved planning, not through better danger avoidance. Three proprietary reasoning models reach notably higher safety awareness (71-81%), while non-reasoning proprietary models and open-source reasoning models remain below 57%. As planning ability approaches saturation for frontier models, improving safety awareness becomes a central challenge for deploying language-model planners in robotic systems.

Landson Guo, Andres M. Diaz Aguilar, William Talbot et al.

Accurate point-wise velocity estimation in 3D is crucial for robot interaction with non-rigid dynamic agents, enabling robust performance in path planning, collision avoidance, and object manipulation in dynamic environments. To this end, this paper proposes a novel RADAR, LiDAR, and camera fusion pipeline for point-wise 3D velocity estimation named CaRLi-V. This pipeline leverages raw RADAR measurements to create a novel RADAR representation, the velocity cube, which densely encodes RADAR radial velocities. By combining the velocity cube for radial velocity extraction, optical flow for tangential velocity estimation, and LiDAR for point-wise range measurements through a closed-form solution, our approach can produce 3D velocity estimates for a dense array of points. Developed as an open-source ROS2 package, CaRLi-V has been field-tested on a custom dataset and achieves low velocity error metrics relative to ground truth while outperforming state-of-the-art scene flow methods.

Gian Erni, Jonas Frey, Takahiro Miki et al.

Elevation maps are commonly used to represent the environment of mobile robots and are instrumental for locomotion and navigation tasks. However, pure geometric information is insufficient for many field applications that require appearance or semantic information, which limits their applicability to other platforms or domains. In this work, we extend a 2.5D robot-centric elevation mapping framework by fusing multi-modal information from multiple sources into a popular map representation. The framework allows inputting data contained in point clouds or images in a unified manner. To manage the different nature of the data, we also present a set of fusion algorithms that can be selected based on the information type and user requirements. Our system is designed to run on the GPU, making it real-time capable for various robotic and learning tasks. We demonstrate the capabilities of our framework by deploying it on multiple robots with varying sensor configurations and showcasing a range of applications that utilize multi-modal layers, including line detection, human detection, and colorization.

Jiaqi Chen, Jonas Frey, Ruyi Zhou et al.

Identifying the physical properties of the surrounding environment is essential for robotic locomotion and navigation to deal with non-geometric hazards, such as slippery and deformable terrains. It would be of great benefit for robots to anticipate these extreme physical properties before contact; however, estimating environmental physical parameters from vision is still an open challenge. Animals can achieve this by using their prior experience and knowledge of what they have seen and how it felt. In this work, we propose a cross-modal self-supervised learning framework for vision-based environmental physical parameter estimation, which paves the way for future physical-property-aware locomotion and navigation. We bridge the gap between existing policies trained in simulation and identification of physical terrain parameters from vision. We propose to train a physical decoder in simulation to predict friction and stiffness from multi-modal input. The trained network allows the labeling of real-world images with physical parameters in a self-supervised manner to further train a visual network during deployment, which can densely predict the friction and stiffness from image data. We validate our physical decoder in simulation and the real world using a quadruped ANYmal robot, outperforming an existing baseline method. We show that our visual network can predict the physical properties in indoor and outdoor experiments while allowing fast adaptation to new environments.

Mike Zhang, Yuntao Ma, Takahiro Miki et al.

Using doors is a longstanding challenge in robotics and is of significant practical interest in giving robots greater access to human-centric spaces. The task is challenging due to the need for online adaptation to varying door properties and precise control in manipulating the door panel and navigating through the confined doorway. To address this, we propose a learning-based controller for a legged manipulator to open and traverse through doors. The controller is trained using a teacher-student approach in simulation to learn robust task behaviors as well as estimate crucial door properties during the interaction. Unlike previous works, our approach is a single control policy that can handle both push and pull doors through learned behaviour which infers the opening direction during deployment without prior knowledge. The policy was deployed on the ANYmal legged robot with an arm and achieved a success rate of 95.0% in repeated trials conducted in an experimental setting. Additional experiments validate the policy's effectiveness and robustness to various doors and disturbances. A video overview of the method and experiments can be found at youtu.be/tQDZXN_k5NU.

Lukas Schneider, Jonas Frey, Takahiro Miki et al.

Deployment in hazardous environments requires robots to understand the risks associated with their actions and movements to prevent accidents. Despite its importance, these risks are not explicitly modeled by currently deployed locomotion controllers for legged robots. In this work, we propose a risk sensitive locomotion training method employing distributional reinforcement learning to consider safety explicitly. Instead of relying on a value expectation, we estimate the complete value distribution to account for uncertainty in the robot's interaction with the environment. The value distribution is consumed by a risk metric to extract risk sensitive value estimates. These are integrated into Proximal Policy Optimization (PPO) to derive our method, Distributional Proximal Policy Optimization (DPPO). The risk preference, ranging from risk-averse to risk-seeking, can be controlled by a single parameter, which enables to adjust the robot's behavior dynamically. Importantly, our approach removes the need for additional reward function tuning to achieve risk sensitivity. We show emergent risk sensitive locomotion behavior in simulation and on the quadrupedal robot ANYmal. Videos of the experiments and code are available at https://sites.google.com/leggedrobotics.com/risk-aware-locomotion.

Zhihao Cao, Tianxu An, Chenhao Li et al.

Collaborative transport requires robots to infer partner intent through physical interaction while maintaining stable loco-manipulation. This becomes particularly challenging in complex environments, where interaction signals are difficult to capture and model. We present PAINT, a lightweight yet efficient hierarchical learning framework for partner-agonistic intent-aware collaborative legged transport that infers partner intent directly from proprioceptive feedback. PAINT decouples intent understanding from terrain-robust locomotion: A high-level policy infers the partner interaction wrench using an intent estimator and a teacher-student training scheme, while a low-level locomotion backbone ensures robust execution. This enables lightweight deployment without external force-torque sensing or payload tracking. Extensive simulation and real-world experiments demonstrate compliant cooperative transport across diverse terrains, payloads, and partners. Furthermore, we show that PAINT naturally scales to decentralized multi-robot transport and transfers across robot embodiments by swapping the underlying locomotion backbone. Our results suggest that proprioceptive signals in payload-coupled interaction provide a scalable interface for partner-agnostic intent-aware collaborative transport.

Manthan Patel, Jonas Frey, Deegan Atha et al.

Autonomous robot navigation in off-road environments requires a comprehensive understanding of the terrain geometry and traversability. The degraded perceptual conditions and sparse geometric information at longer ranges make the problem challenging especially when driving at high speeds. Furthermore, the sensing-to-mapping latency and the look-ahead map range can limit the maximum speed of the vehicle. Building on top of the recent work RoadRunner, in this work, we address the challenge of long-range (100 m) traversability estimation. Our RoadRunner (M&M) is an end-to-end learning-based framework that directly predicts the traversability and elevation maps at multiple ranges (50 m, 100 m) and resolutions (0.2 m, 0.8 m) taking as input multiple images and a LiDAR voxel map. Our method is trained in a self-supervised manner by leveraging the dense supervision signal generated by fusing predictions from an existing traversability estimation stack (X-Racer) in hindsight and satellite Digital Elevation Maps. RoadRunner M&M achieves a significant improvement of up to 50% for elevation mapping and 30% for traversability estimation over RoadRunner, and is able to predict in 30% more regions compared to X-Racer while achieving real-time performance. Experiments on various out-of-distribution datasets also demonstrate that our data-driven approach starts to generalize to novel unstructured environments. We integrate our proposed framework in closed-loop with the path planner to demonstrate autonomous high-speed off-road robotic navigation in challenging real-world environments. Project Page: https://leggedrobotics.github.io/roadrunner_mm/

Jin Jin, Chong Zhang, Jonas Frey et al.

Autonomous robots must navigate reliably in unknown environments even under compromised exteroceptive perception, or perception failures. Such failures often occur when harsh environments lead to degraded sensing, or when the perception algorithm misinterprets the scene due to limited generalization. In this paper, we model perception failures as invisible obstacles and pits, and train a reinforcement learning (RL) based local navigation policy to guide our legged robot. Unlike previous works relying on heuristics and anomaly detection to update navigational information, we train our navigation policy to reconstruct the environment information in the latent space from corrupted perception and react to perception failures end-to-end. To this end, we incorporate both proprioception and exteroception into our policy inputs, thereby enabling the policy to sense collisions on different body parts and pits, prompting corresponding reactions. We validate our approach in simulation and on the real quadruped robot ANYmal running in real-time (<10 ms CPU inference). In a quantitative comparison with existing heuristic-based locally reactive planners, our policy increases the success rate over 30% when facing perception failures. Project Page: https://bit.ly/45NBTuh.

René Zurbrügg, Tifanny Portela, Arjun Bhardwaj et al.

Learning from demonstrations is effective for robotic manipulation, but collecting sufficient task-specific data remains a major bottleneck. Under distribution shift, small errors compound, performance degrades, and expert time is often spent on redundant, low-value corrections instead of the few critical failure cases.

Guido Caccianiga, Julian Nubert, Cesar Cadena et al.

Reconstructing the 3D shape of a deformable environment from the information captured by a moving depth camera is highly relevant to surgery. The underlying challenge is the fact that simultaneously estimating camera motion and tissue deformation in a fully deformable scene is an ill-posed problem, especially from a single arbitrarily moving viewpoint. Current solutions are often organ-specific and lack the robustness required to handle large deformations. Here we propose a multi-viewpoint global optimization framework that can flexibly integrate the output of low-level perception modules (data association, depth, and relative scene flow) with kinematic and scene-modeling priors to jointly estimate multiple camera motions and absolute scene flow. We use simulated noisy data to show three practical examples that successfully constrain the convergence to a unique solution. Overall, our method shows robustness to combined noisy input measures and can process hundreds of points in a few milliseconds. MultiViPerFrOG builds a generalized learning-free scaffolding for spatio-temporal encoding that can unlock advanced surgical scene representations and will facilitate the development of the computer-assisted-surgery technologies of the future.

Gianluca Sabatini, Chenhao Li, Marco Hutter

Proximal Policy Optimization (PPO) has become the de facto standard for training legged robots, thanks to its robustness and scalability in massively parallel simulation environments like IsaacLab. However, its on-policy nature makes it inherently sample-inefficient, preventing its use for continuous adaptation and fine-tuning on real hardware. Soft Actor-Critic (SAC), by contrast, is an off-policy algorithm that can reuse past experience, making it a natural candidate for sim-to-real transfer workflows where the same algorithm can be used both in simulation and for online learning on the real robot. Despite these advantages, SAC has consistently failed to match PPO's empirical performance in massively parallel training settings. This work identifies the root causes of this gap and introduces targeted modifications, covering policy initialization, timeout-aware critic targets, and multi-step return estimation, that enable SAC to train stably at scale. Evaluated across multiple legged robot platforms and diverse locomotion tasks, our approach closes the performance gap with PPO entirely.

Tifanny Portela, Andrei Cramariuc, Mayank Mittal et al.

Combining manipulation with the mobility of legged robots is essential for a wide range of robotic applications. However, integrating an arm with a mobile base significantly increases the system's complexity, making precise end-effector control challenging. Existing model-based approaches are often constrained by their modeling assumptions, leading to limited robustness. Meanwhile, recent Reinforcement Learning (RL) implementations restrict the arm's workspace to be in front of the robot or track only the position to obtain decent tracking accuracy. In this work, we address these limitations by introducing a whole-body RL formulation for end-effector pose tracking in a large workspace on rough, unstructured terrains. Our proposed method involves a terrain-aware sampling strategy for the robot's initial configuration and end-effector pose commands, as well as a game-based curriculum to extend the robot's operating range. We validate our approach on the ANYmal quadrupedal robot with a six DoF robotic arm. Through our experiments, we show that the learned controller achieves precise command tracking over a large workspace and adapts across varying terrains such as stairs and slopes. On deployment, it achieves a pose-tracking error of 2.64 cm and 3.64 degrees, outperforming existing competitive baselines.

Hardik Shah, Erica Tevere, Deegan Atha et al.

Autonomous navigation in complex, unstructured outdoor environments requires robots to operate over long ranges without prior maps and limited depth sensing. In such settings, relying solely on geometric frontiers for exploration is often insufficient. In such settings, the ability to reason semantically about where to go and what is safe to traverse is crucial for robust, efficient exploration. This work presents WildOS, a unified system for long-range, open-vocabulary object search that combines safe geometric exploration with semantic visual reasoning. WildOS builds a sparse navigation graph to maintain spatial memory, while utilizing a foundation-model-based vision module, ExploRFM, to score frontier nodes of the graph. ExploRFM simultaneously predicts traversability, visual frontiers, and object similarity in image space, enabling real-time, onboard semantic navigation tasks. The resulting vision-scored graph enables the robot to explore semantically meaningful directions while ensuring geometric safety. Furthermore, we introduce a particle-filter-based method for coarse localization of the open-vocabulary target query, that estimates candidate goal positions beyond the robot's immediate depth horizon, enabling effective planning toward distant goals. Extensive closed-loop field experiments across diverse off-road and urban terrains demonstrate that WildOS enables robust navigation, significantly outperforming purely geometric and purely vision-based baselines in both efficiency and autonomy. Our results highlight the potential of vision foundation models to drive open-world robotic behaviors that are both semantically informed and geometrically grounded. Project Page: https://leggedrobotics.github.io/wildos/

Kaixian Qu, Han Wang, Victor Klemm et al.

Active perception is a fundamental problem in autonomous robotics in which the robot must decide where to move and what to sense in order to obtain the most informative observations for accomplishing its mission. Existing approaches either solve a computationally expensive traveling salesman problem over heuristically selected informative nodes, or adopt a more efficient but overly constrained shortest path tree formulation. To address these limitations, we explore beam search algorithms as scalable alternatives. While the standard beam search provides scalability by preserving the top-B paths at each depth level, it is prone to local optima and exhibits parameter sensitivity. Our first contribution is a node-wise beam search (NBS) algorithm, which maintains top-B candidates per node to enable more effective exploration of the solution space. Systematic benchmarking on graphs shows that NBS consistently outperforms other baselines and maintains strong performance even at low beam widths. As a second contribution, we integrate the concept of frontiers into the path selection criterion, introducing the expected gain metric, which better balances exploration and exploitation compared to existing alternatives. Our third contribution proposes the rapidly-exploring random annulus graph (RRAG), a novel graph construction method that preserves full orientation sampling and ensures connectivity in cluttered environments through a fallback local sampling-based planner. Extensive experiments demonstrate that NBS combined with RRAG achieves the highest performance across all three representative active perception tasks, outperforming state-of-the-art algorithms by at least 20% in one or more tasks. We further validate the approach on real robotic platforms in different scenarios.

Zhengyu Fu, René Zurbrügg, Kaixian Qu et al.

Recent work in 3D scene understanding is moving beyond purely spatial analysis toward functional scene understanding. However, existing methods often consider functional relationships between object pairs in isolation, failing to capture the scene-wide interdependence that humans use to resolve ambiguity. We introduce FunFact, a framework for constructing probabilistic open-vocabulary functional 3D scene graphs from posed RGB-D images. FunFact first builds an object- and part-centric 3D map and uses foundation models to propose semantically plausible functional relations. These candidates are converted into factor graph variables and constrained by both LLM-derived common-sense priors and geometric priors. This formulation enables joint probabilistic inference over all functional edges and their marginals, yielding substantially better calibrated confidence scores. To benchmark this setting, we introduce FunThor, a synthetic dataset based on AI2-THOR with part-level geometry and rule-based functional annotations. Experiments on SceneFun3D, FunGraph3D, and FunThor show that FunFact improves node and relation discovery recall and significantly reduces calibration error for ambiguous relations, highlighting the benefits of holistic probabilistic modeling for functional scene understanding. See our project page at https://funfact-scenegraph.github.io/

Mike Zhang, Kaixian Qu, Vaishakh Patil et al.

Large Language Models (LLM) have emerged as a tool for robots to generate task plans using common sense reasoning. For the LLM to generate actionable plans, scene context must be provided, often through a map. Recent works have shifted from explicit maps with fixed semantic classes to implicit open vocabulary maps based on queryable embeddings capable of representing any semantic class. However, embeddings cannot directly report the scene context as they are implicit, requiring further processing for LLM integration. To address this, we propose an explicit text-based map that can represent thousands of semantic classes while easily integrating with LLMs due to their text-based nature by building upon large-scale image recognition models. We study how entities in our map can be localized and show through evaluations that our text-based map localizations perform comparably to those from open vocabulary maps while using two to four orders of magnitude less memory. Real-robot experiments demonstrate the grounding of an LLM with the text-based map to solve user tasks.

Arjun Bhardwaj, Maximum Wilder-Smith, Mayank Mittal et al.

In-hand object reorientation requires precise estimation of the object pose to handle complex task dynamics. While RGB sensing offers rich semantic cues for pose tracking, existing solutions rely on multi-camera setups or costly ray tracing. We present a sim-to-real framework for monocular RGB in-hand reorientation that integrates 3D Gaussian Splatting (3DGS) to bridge the visual sim-to-real gap. Our key insight is performing domain randomization in the Gaussian representation space: by applying physically consistent, pre-rendering augmentations to 3D Gaussians, we generate photorealistic, randomized visual data for object pose estimation. The manipulation policy is trained using curriculum-based reinforcement learning with teacher-student distillation, enabling efficient learning of complex behaviors. Importantly, both perception and control models can be trained independently on consumer-grade hardware, eliminating the need for large compute clusters. Experiments show that the pose estimator trained with 3DGS data outperforms those trained using conventional rendering data in challenging visual environments. We validate the system on a physical multi-fingered hand equipped with an RGB camera, demonstrating robust reorientation of five diverse objects even under challenging lighting conditions. Our results highlight Gaussian splatting as a practical path for RGB-only dexterous manipulation. For videos of the hardware deployments and additional supplementary materials, please refer to the project website: https://rffr.leggedrobotics.com/works/viserdex/

Chong Zhang, Victor Klemm, Fan Yang et al.

Achieving agile and generalized legged locomotion across terrains requires tight integration of perception and control, especially under occlusions and sparse footholds. Existing methods have demonstrated agility on parkour courses but often rely on end-to-end sensorimotor models with limited generalization and interpretability. By contrast, methods targeting generalized locomotion typically exhibit limited agility and struggle with visual occlusions. We introduce AME-2, a unified reinforcement learning (RL) framework for agile and generalized locomotion that incorporates a novel attention-based map encoder in the control policy. This encoder extracts local and global mapping features and uses attention mechanisms to focus on salient regions, producing an interpretable and generalized embedding for RL-based control. We further propose a learning-based mapping pipeline that provides fast, uncertainty-aware terrain representations robust to noise and occlusions, serving as policy inputs. It uses neural networks to convert depth observations into local elevations with uncertainties, and fuses them with odometry. The pipeline also integrates with parallel simulation so that we can train controllers with online mapping, aiding sim-to-real transfer. We validate AME-2 with the proposed mapping pipeline on a quadruped and a biped robot, and the resulting controllers demonstrate strong agility and generalization to unseen terrains in simulation and in real-world experiments.

Philipp Schoch, Fan Yang, Yuntao Ma et al.

Current visual navigation systems often treat the environment as static, lacking the ability to adaptively interact with obstacles. This limitation leads to navigation failure when encountering unavoidable obstructions. In response, we introduce IN-Sight, a novel approach to self-supervised path planning, enabling more effective navigation strategies through interaction with obstacles. Utilizing RGB-D observations, IN-Sight calculates traversability scores and incorporates them into a semantic map, facilitating long-range path planning in complex, maze-like environments. To precisely navigate around obstacles, IN-Sight employs a local planner, trained imperatively on a differentiable costmap using representation learning techniques. The entire framework undergoes end-to-end training within the state-of-the-art photorealistic Intel SPEAR Simulator. We validate the effectiveness of IN-Sight through extensive benchmarking in a variety of simulated scenarios and ablation studies. Moreover, we demonstrate the system's real-world applicability with zero-shot sim-to-real transfer, deploying our planner on the legged robot platform ANYmal, showcasing its practical potential for interactive navigation in real environments.

Zewei Zhang, Kehan Wen, Michael Xu et al.

Whole-body humanoid locomotion is challenging due to high-dimensional control, morphological instability, and the need for real-time adaptation to various terrains using onboard perception. Directly applying reinforcement learning (RL) with reward shaping to humanoid locomotion often leads to lower-body-dominated behaviors, whereas imitation-based RL can learn more coordinated whole-body skills but is typically limited to replaying reference motions without a mechanism to adapt them online from perception for terrain-aware locomotion. To address this gap, we propose a whole-body humanoid locomotion framework that combines skills learned from reference motions with terrain-aware adaptation. We first train a diffusion model on retargeted human motions for real-time prediction of terrain-aware reference motions. Concurrently, we train a whole-body reference tracker with RL using this motion data. To improve robustness under imperfectly generated references, we further fine-tune the tracker with a frozen motion generator in a closed-loop setting. The resulting system supports directional goal-reaching control with terrain-aware whole-body adaptation, and can be deployed on a Unitree G1 humanoid robot with onboard perception and computation. The hardware experiments demonstrate successful traversal over boxes, hurdles, stairs, and mixed terrain combinations. Quantitative results further show the benefits of incorporating online motion generation and fine-tuning the motion tracker for improved generalization and robustness.

Francesca Bray, Simone Tolomei, Andrei Cramariuc et al.

Robotic collaborative carrying could greatly benefit human activities like warehouse and construction site management. However, coordinating the simultaneous motion of multiple robots represents a significant challenge. Existing works primarily focus on obstacle-free environments, making them unsuitable for most real-world applications. Works that account for obstacles, either overfit to a specific terrain configuration or rely on pre-recorded maps combined with path planners to compute collision-free trajectories. This work focuses on two quadrupedal robots mechanically connected to a carried object. We propose a Reinforcement Learning (RL)-based policy that enables tracking a commanded velocity direction while avoiding collisions with nearby obstacles using only onboard sensing, eliminating the need for precomputed trajectories and complete map knowledge. Our work presents a hierarchical architecture, where a perceptive high-level object-centric policy commands two pretrained locomotion policies. Additionally, we employ a game-inspired curriculum to increase the complexity of obstacles in the terrain progressively. We validate our approach on two quadrupedal robots connected to a bar via spherical joints, benchmarking it against optimization-based and decentralized RL baselines. Our hardware experiments demonstrate the ability of our system to locomote in unknown environments without the need for a map or a path planner. The video of our work is available in the multimedia material.

Carmen Scheidemann, Andrei Cramariuc, Changan Chen et al.

Background: Assistance robots have the potential to increase the independence of people who need daily care due to limited mobility or being wheelchair-bound. Current solutions of attaching robotic arms to motorized wheelchairs offer limited additional mobility at the cost of increased size and reduced wheelchair maneuverability. Methods: We present an on-demand quadrupedal assistance robot system controlled via a shared autonomy approach, which combines semi-autonomous task execution with human teleoperation. Due to the mobile nature of the system it can assist the operator whenever needed and perform autonomous tasks independently, without otherwise restricting their mobility. We automate pick-and-place tasks, as well as robot movement through the environment with semantic, collision-aware navigation. For teleoperation, we present a mouth-level joystick interface that enables an operator with reduced mobility to control the robot's end effector for precision manipulation. Results: We showcase our system in the \textit{Cybathlon 2024 Assistance Robot Race}, and validate it in an at-home experimental setup, where we measure task completion times and user satisfaction. We find our system capable of assisting in a broad variety of tasks, including those that require dexterous manipulation. The user study confirms the intuition that increased robot autonomy alleviates the operator's mental load. Conclusions: We present a flexible system that has the potential to help people in wheelchairs maintain independence in everyday life by enabling them to solve mobile manipulation problems without external support. We achieve results comparable to previous state-of-the-art on subjective metrics while allowing for more autonomy of the operator and greater agility for manipulation.

Florentin Dipner, William Talbot, Turcan Tuna et al.

This paper presents a robust one-shot badminton shuttlecock detection framework for non-stationary robots. To address the lack of egocentric shuttlecock detection datasets, we introduce a dataset of 20,510 semi-automatically annotated frames captured across 11 distinct backgrounds in diverse indoor and outdoor environments, and categorize each frame into one of three difficulty levels. For labeling, we present a novel semi-automatic annotation pipeline, that enables efficient labeling from stationary camera footage. We propose a metric suited to our downstream use case and fine-tune a YOLOv8 network optimized for real-time shuttlecock detection, achieving an F1-score of 0.86 under our metric in test environments similar to training, and 0.70 in entirely unseen environments. Our analysis reveals that detection performance is critically dependent on shuttlecock size and background texture complexity. Qualitative experiments confirm their applicability to robots with moving cameras. Unlike prior work with stationary camera setups, our detector is specifically designed for the egocentric, dynamic viewpoints of mobile robots, providing a foundational building block for downstream tasks, including tracking, trajectory estimation, and system (re)-initialization.

Julian Nubert, Turcan Tuna, Jonas Frey et al.

Seamless operation of mobile robots in challenging environments requires low-latency local motion estimation (e.g., dynamic maneuvers) and accurate global localization (e.g., wayfinding). While most existing sensor-fusion approaches are designed for specific scenarios, this work introduces a flexible open-source solution for task- and setup-agnostic multimodal sensor fusion that is distinguished by its generality and usability. Holistic Fusion formulates sensor fusion as a combined estimation problem of i) the local and global robot state and ii) a (theoretically unlimited) number of dynamic context variables, including automatic alignment of reference frames; this formulation fits countless real-world applications without any conceptual modifications. The proposed factor-graph solution enables the direct fusion of an arbitrary number of absolute, local, and landmark measurements expressed with respect to different reference frames by explicitly including them as states in the optimization and modeling their evolution as random walks. Moreover, local smoothness and consistency receive particular attention to prevent jumps in the robot state belief. HF enables low-latency and smooth online state estimation on typical robot hardware while simultaneously providing low-drift global localization at the IMU measurement rate. The efficacy of this released framework is demonstrated in five real-world scenarios on three robotic platforms, each with distinct task requirements.

Clemens Schwarke, Mayank Mittal, Nikita Rudin et al.

RSL-RL is an open-source Reinforcement Learning library tailored to the specific needs of the robotics community. Unlike broad general-purpose frameworks, its design philosophy prioritizes a compact and easily modifiable codebase, allowing researchers to adapt and extend algorithms with minimal overhead. The library focuses on algorithms most widely adopted in robotics, together with auxiliary techniques that address robotics-specific challenges. Optimized for GPU-only training, RSL-RL achieves high-throughput performance in large-scale simulation environments. Its effectiveness has been validated in both simulation benchmarks and in real-world robotic experiments, demonstrating its utility as a lightweight, extensible, and practical framework to develop learning-based robotic controllers. The library is open-sourced at: https://github.com/leggedrobotics/rsl_rl.

Mateo Haro, Julia Richter, Fan Yang et al.

Long-range navigation is commonly addressed through hierarchical pipelines in which a global planner generates a path, decomposed into waypoints, and followed sequentially by a local planner. These systems are sensitive to global path quality, as inaccurate remote sensing data can result in locally infeasible waypoints, which degrade local execution. At the same time, the limited global context available to the local planner hinders long-range efficiency. To address this issue, we propose a reinforcement learning-based local navigation policy that leverages path information as contextual guidance. The policy is conditioned on reference path observations and trained with a reward function mainly based on goal-reaching objectives, without any explicit path-following reward. Through this implicit conditioning, the policy learns to opportunistically exploit path information while remaining robust to misleading or degraded guidance. Experimental results show that the proposed approach significantly improves navigation efficiency when high-quality paths are available and maintains baseline-level performance when path observations are severely degraded or even non-existent. These properties make the method particularly well-suited for long-range navigation scenarios in which high-level plans are approximate and local execution must remain adaptive to uncertainty.

Svitlana Morkva, Maximum Wilder-Smith, Michael Oechsle et al.

We present MOSAIC-GS, a novel, fully explicit, and computationally efficient approach for high-fidelity dynamic scene reconstruction from monocular videos using Gaussian Splatting. Monocular reconstruction is inherently ill-posed due to the lack of sufficient multiview constraints, making accurate recovery of object geometry and temporal coherence particularly challenging. To address this, we leverage multiple geometric cues, such as depth, optical flow, dynamic object segmentation, and point tracking. Combined with rigidity-based motion constraints, these cues allow us to estimate preliminary 3D scene dynamics during an initialization stage. Recovering scene dynamics prior to the photometric optimization reduces reliance on motion inference from visual appearance alone, which is often ambiguous in monocular settings. To enable compact representations, fast training, and real-time rendering while supporting non-rigid deformations, the scene is decomposed into static and dynamic components. Each Gaussian in the dynamic part of the scene is assigned a trajectory represented as time-dependent Poly-Fourier curve for parameter-efficient motion encoding. We demonstrate that MOSAIC-GS achieves substantially faster optimization and rendering compared to existing methods, while maintaining reconstruction quality on par with state-of-the-art approaches across standard monocular dynamic scene benchmarks.

Mohamad H. Danesh, Chenhao Li, Amin Abyaneh et al.

World models promise a paradigm shift in robotics, where an agent learns the underlying physics of its environment once to enable efficient planning and behavior learning. However, current world models are often hardware-locked specialists: a model trained on a Boston Dynamics Spot robot fails catastrophically on a Unitree Go1 due to the mismatch in kinematic and dynamic properties, as the model overfits to specific embodiment constraints rather than capturing the universal locomotion dynamics. Consequently, a slight change in actuator dynamics or limb length necessitates training a new model from scratch. In this work, we take a step towards a framework for training a generalizable Quadrupedal World Model (QWM) that disentangles environmental dynamics from robot morphology. We address the limitations of implicit system identification, where treating static physical properties (like mass or limb length) as latent variables to be inferred from motion history creates an adaptation lag that can compromise zero-shot safety and efficiency. Instead, we explicitly condition the generative dynamics on the robot's engineering specifications. By integrating a physical morphology encoder and a reward normalizer, we enable the model to serve as a neural simulator capable of generalizing across morphologies. This capability unlocks zero-shot control across a range of embodiments. We introduce, for the first time, a world model that enables zero-shot generalization to new morphologies for locomotion. While we carefully study the limitations of our method, QWM operates as a distribution-bounded interpolator within the quadrupedal morphology family rather than a universal physics engine, this work represents a significant step toward morphology-conditioned world models for legged locomotion.

Lennart Werner, Pol Eyschen, Sean Costello et al.

High-precision heavy-duty grading is a common step in earthworks, traditionally carried out manually by skilled operators. Removing a significant amount of material while achieving a high-precision surface requires substantial machine-specific experience. Different hydraulic architectures react differently to operator inputs and soil interaction forces, which makes generalizable controllers challenging. In this paper, we present an autonomous controller that achieves high-precision grading at expert-operator speed on Load Sensing and Negative Flow Control machines alike. We split our controller into two parts: (1) a hydraulic-aware low-level loop that is hydraulic architecture-specific and (2) a path-tracking layer that coordinates joint motions and responses. Through a calibration process, our technique is applicable to load-sensing and negative-flow-control machinery. To showcase its versatility, we benchmark our approach on two excavators with different hydraulics and compare it against a commercial state-of-the-art solution. Our technique (RMSE 1.8~cm) outperforms the commercial solution (RMSE 4.7~cm) in precision by a factor of 2.6 and improves machine usage by leveraging the maximum function pressure, as opposed to commercial solutions that stall prematurely.

Christopher E. Mower, Yuhui Wan, Hongzhan Yu et al.

We present a framework for intuitive robot programming by non-experts, leveraging natural language prompts and contextual information from the Robot Operating System (ROS). Our system integrates large language models (LLMs), enabling non-experts to articulate task requirements to the system through a chat interface. Key features of the framework include: integration of ROS with an AI agent connected to a plethora of open-source and commercial LLMs, automatic extraction of a behavior from the LLM output and execution of ROS actions/services, support for three behavior modes (sequence, behavior tree, state machine), imitation learning for adding new robot actions to the library of possible actions, and LLM reflection via human and environment feedback. Extensive experiments validate the framework, showcasing robustness, scalability, and versatility in diverse scenarios, including long-horizon tasks, tabletop rearrangements, and remote supervisory control. To facilitate the adoption of our framework and support the reproduction of our results, we have made our code open-source. You can access it at: https://github.com/huawei-noah/HEBO/tree/master/ROSLLM.

Nikita Rudin, David Hoeller, Philipp Reist et al.

In this work, we present and study a training set-up that achieves fast policy generation for real-world robotic tasks by using massive parallelism on a single workstation GPU. We analyze and discuss the impact of different training algorithm components in the massively parallel regime on the final policy performance and training times. In addition, we present a novel game-inspired curriculum that is well suited for training with thousands of simulated robots in parallel. We evaluate the approach by training the quadrupedal robot ANYmal to walk on challenging terrain. The parallel approach allows training policies for flat terrain in under four minutes, and in twenty minutes for uneven terrain. This represents a speedup of multiple orders of magnitude compared to previous work. Finally, we transfer the policies to the real robot to validate the approach. We open-source our training code to help accelerate further research in the field of learned legged locomotion.

Joonho Lee, Marko Bjelonic, Alexander Reske et al.

Autonomous wheeled-legged robots have the potential to transform logistics systems, improving operational efficiency and adaptability in urban environments. Navigating urban environments, however, poses unique challenges for robots, necessitating innovative solutions for locomotion and navigation. These challenges include the need for adaptive locomotion across varied terrains and the ability to navigate efficiently around complex dynamic obstacles. This work introduces a fully integrated system comprising adaptive locomotion control, mobility-aware local navigation planning, and large-scale path planning within the city. Using model-free reinforcement learning (RL) techniques and privileged learning, we develop a versatile locomotion controller. This controller achieves efficient and robust locomotion over various rough terrains, facilitated by smooth transitions between walking and driving modes. It is tightly integrated with a learned navigation controller through a hierarchical RL framework, enabling effective navigation through challenging terrain and various obstacles at high speed. Our controllers are integrated into a large-scale urban navigation system and validated by autonomous, kilometer-scale navigation missions conducted in Zurich, Switzerland, and Seville, Spain. These missions demonstrate the system's robustness and adaptability, underscoring the importance of integrated control systems in achieving seamless navigation in complex environments. Our findings support the feasibility of wheeled-legged robots and hierarchical RL for autonomous navigation, with implications for last-mile delivery and beyond.

Philip Arm, Mayank Mittal, Hendrik Kolvenbach et al.

Legged robots have the potential to become vital in maintenance, home support, and exploration scenarios. In order to interact with and manipulate their environments, most legged robots are equipped with a dedicated robot arm, which means additional mass and mechanical complexity compared to standard legged robots. In this work, we explore pedipulation - using the legs of a legged robot for manipulation. By training a reinforcement learning policy that tracks position targets for one foot, we enable a dedicated pedipulation controller that is robust to disturbances, has a large workspace through whole-body behaviors, and can reach far-away targets with gait emergence, enabling loco-pedipulation. By deploying our controller on a quadrupedal robot using teleoperation, we demonstrate various real-world tasks such as door opening, sample collection, and pushing obstacles. We demonstrate load carrying of more than 2.0 kg at the foot. Additionally, the controller is robust to interaction forces at the foot, disturbances at the base, and slippery contact surfaces. Videos of the experiments are available at https://sites.google.com/leggedrobotics.com/pedipulate.