Karl Van Wyk

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.

Ankur Handa, Arthur Allshire, Viktor Makoviychuk et al. · cmu

Recent work has demonstrated the ability of deep reinforcement learning (RL) algorithms to learn complex robotic behaviours in simulation, including in the domain of multi-fingered manipulation. However, such models can be challenging to transfer to the real world due to the gap between simulation and reality. In this paper, we present our techniques to train a) a policy that can perform robust dexterous manipulation on an anthropomorphic robot hand and b) a robust pose estimator suitable for providing reliable real-time information on the state of the object being manipulated. Our policies are trained to adapt to a wide range of conditions in simulation. Consequently, our vision-based policies significantly outperform the best vision policies in the literature on the same reorientation task and are competitive with policies that are given privileged state information via motion capture systems. Our work reaffirms the possibilities of sim-to-real transfer for dexterous manipulation in diverse kinds of hardware and simulator setups, and in our case, with the Allegro Hand and Isaac Gym GPU-based simulation. Furthermore, it opens up possibilities for researchers to achieve such results with commonly-available, affordable robot hands and cameras. Videos of the resulting policy and supplementary information, including experiments and demos, can be found at https://dextreme.org/

Yuzhe Qin, Wei Yang, Binghao Huang et al. · nvidia

Vision-based teleoperation offers the possibility to endow robots with human-level intelligence to physically interact with the environment, while only requiring low-cost camera sensors. However, current vision-based teleoperation systems are designed and engineered towards a particular robot model and deploy environment, which scales poorly as the pool of the robot models expands and the variety of the operating environment increases. In this paper, we propose AnyTeleop, a unified and general teleoperation system to support multiple different arms, hands, realities, and camera configurations within a single system. Although being designed to provide great flexibility to the choice of simulators and real hardware, our system can still achieve great performance. For real-world experiments, AnyTeleop can outperform a previous system that was designed for a specific robot hardware with a higher success rate, using the same robot. For teleoperation in simulation, AnyTeleop leads to better imitation learning performance, compared with a previous system that is particularly designed for that simulator. Project page: https://yzqin.github.io/anyteleop/.

Zoey Qiuyu Chen, Karl Van Wyk, Yu-Wei Chao et al. · nvidia

Teaching a multi-fingered dexterous robot to grasp objects in the real world has been a challenging problem due to its high dimensional state and action space. We propose a robot-learning system that can take a small number of human demonstrations and learn to grasp unseen object poses given partially occluded observations. Our system leverages a small motion capture dataset and generates a large dataset with diverse and successful trajectories for a multi-fingered robot gripper. By adding domain randomization, we show that our dataset provides robust grasping trajectories that can be transferred to a policy learner. We train a dexterous grasping policy that takes the point clouds of the object as input and predicts continuous actions to grasp objects from different initial robot states. We evaluate the effectiveness of our system on a 22-DoF floating Allegro Hand in simulation and a 23-DoF Allegro robot hand with a KUKA arm in real world. The policy learned from our dataset can generalize well on unseen object poses in both simulation and the real world

Ritvik Singh, Jingzhou Liu, Karl Van Wyk et al.

Vision-based object detectors are a crucial basis for robotics applications as they provide valuable information about object localisation in the environment. These need to ensure high reliability in different lighting conditions, occlusions, and visual artifacts, all while running in real-time. Collecting and annotating real-world data for these networks is prohibitively time consuming and costly, especially for custom assets, such as industrial objects, making it untenable for generalization to in-the-wild scenarios. To this end, we present Synthetica, a method for large-scale synthetic data generation for training robust state estimators. This paper focuses on the task of object detection, an important problem which can serve as the front-end for most state estimation problems, such as pose estimation. Leveraging data from a photorealistic ray-tracing renderer, we scale up data generation, generating 2.7 million images, to train highly accurate real-time detection transformers. We present a collection of rendering randomization and training-time data augmentation techniques conducive to robust sim-to-real performance for vision tasks. We demonstrate state-of-the-art performance on the task of object detection while having detectors that run at 50-100Hz which is 9 times faster than the prior SOTA. We further demonstrate the usefulness of our training methodology for robotics applications by showcasing a pipeline for use in the real world with custom objects for which there do not exist prior datasets. Our work highlights the importance of scaling synthetic data generation for robust sim-to-real transfer while achieving the fastest real-time inference speeds. Videos and supplementary information can be found at this URL: https://sites.google.com/view/synthetica-vision.

Ritvik Singh, Karl Van Wyk, Pieter Abbeel et al.

This work explores techniques to scale up image-based end-to-end learning for dexterous grasping with an arm + hand system. Unlike state-based RL, vision-based RL is much more memory inefficient, resulting in relatively low batch sizes, which is not amenable for algorithms like PPO. Nevertheless, it is still an attractive method as unlike the more commonly used techniques which distill state-based policies into vision networks, end-to-end RL can allow for emergent active vision behaviors. We identify a key bottleneck in training these policies is the way most existing simulators scale to multiple GPUs using traditional data parallelism techniques. We propose a new method where we disaggregate the simulator and RL (both training and experience buffers) onto separate GPUs. On a node with four GPUs, we have the simulator running on three of them, and PPO running on the fourth. We are able to show that with the same number of GPUs, we can double the number of existing environments compared to the previous baseline of standard data parallelism. This allows us to train vision-based environments, end-to-end with depth, which were previously performing far worse with the baseline. We train and distill both depth and state-based policies into stereo RGB networks and show that depth distillation leads to better results, both in simulation and reality. This improvement is likely due to the observability gap between state and vision policies which does not exist when distilling depth policies into stereo RGB. We further show that the increased batch size brought about by disaggregated simulation also improves real world performance. When deploying in the real world, we improve upon the previous state-of-the-art vision-based results using our end-to-end policies.

Karl Van Wyk, Mandy Xie, Anqi Li et al.

Classical mechanical systems are central to controller design in energy shaping methods of geometric control. However, their expressivity is limited by position-only metrics and the intimate link between metric and geometry. Recent work on Riemannian Motion Policies (RMPs) has shown that shedding these restrictions results in powerful design tools, but at the expense of theoretical stability guarantees. In this work, we generalize classical mechanics to what we call geometric fabrics, whose expressivity and theory enable the design of systems that outperform RMPs in practice. Geometric fabrics strictly generalize classical mechanics forming a new physics of behavior by first generalizing them to Finsler geometries and then explicitly bending them to shape their behavior while maintaining stability. We develop the theory of fabrics and present both a collection of controlled experiments examining their theoretical properties and a set of robot system experiments showing improved performance over a well-engineered and hardened implementation of RMPs, our current state-of-the-art in controller design.

Mandy Xie, Anqi Li, Karl Van Wyk et al.

Imitation learning (IL) is a frequently used approach for data-efficient policy learning. Many IL methods, such as Dataset Aggregation (DAgger), combat challenges like distributional shift by interacting with oracular experts. Unfortunately, assuming access to oracular experts is often unrealistic in practice; data used in IL frequently comes from offline processes such as lead-through or teleoperation. In this paper, we present a novel imitation learning technique called Collocation for Demonstration Encoding (CoDE) that operates on only a fixed set of trajectory demonstrations. We circumvent challenges with methods like back-propagation-through-time by introducing an auxiliary trajectory network, which takes inspiration from collocation techniques in optimal control. Our method generalizes well and more accurately reproduces the demonstrated behavior with fewer guiding trajectories when compared to standard behavioral cloning methods. We present simulation results on a 7-degree-of-freedom (DoF) robotic manipulator that learns to exhibit lifting, target-reaching, and obstacle avoidance behaviors.

Yu-Wei Chao, Wei Yang, Yu Xiang et al.

We introduce DexYCB, a new dataset for capturing hand grasping of objects. We first compare DexYCB with a related one through cross-dataset evaluation. We then present a thorough benchmark of state-of-the-art approaches on three relevant tasks: 2D object and keypoint detection, 6D object pose estimation, and 3D hand pose estimation. Finally, we evaluate a new robotics-relevant task: generating safe robot grasps in human-to-robot object handover. Dataset and code are available at https://dex-ycb.github.io.

Anqi Li, Ching-An Cheng, M. Asif Rana et al.

We consider the problem of learning motion policies for acceleration-based robotics systems with a structured policy class specified by RMPflow. RMPflow is a multi-task control framework that has been successfully applied in many robotics problems. Using RMPflow as a structured policy class in learning has several benefits, such as sufficient expressiveness, the flexibility to inject different levels of prior knowledge as well as the ability to transfer policies between robots. However, implementing a system for end-to-end learning RMPflow policies faces several computational challenges. In this work, we re-examine the message passing algorithm of RMPflow and propose a more efficient alternate algorithm, called RMP2, that uses modern automatic differentiation tools (such as TensorFlow and PyTorch) to compute RMPflow policies. Our new design retains the strengths of RMPflow while bringing in advantages from automatic differentiation, including 1) easy programming interfaces to designing complex transformations; 2) support of general directed acyclic graph (DAG) transformation structures; 3) end-to-end differentiability for policy learning; 4) improved computational efficiency. Because of these features, RMP2 can be treated as a structured policy class for efficient robot learning which is suitable encoding domain knowledge. Our experiments show that using structured policy class given by RMP2 can improve policy performance and safety in reinforcement learning tasks for goal reaching in cluttered space.

Yashraj S. Narang, Balakumar Sundaralingam, Karl Van Wyk et al.

In the human hand, high-density contact information provided by afferent neurons is essential for many human grasping and manipulation capabilities. In contrast, robotic tactile sensors, including the state-of-the-art SynTouch BioTac, are typically used to provide low-density contact information, such as contact location, center of pressure, and net force. Although useful, these data do not convey or leverage the rich information content that some tactile sensors naturally measure. This research extends robotic tactile sensing beyond reduced-order models through 1) the automated creation of a precise experimental tactile dataset for the BioTac over a diverse range of physical interactions, 2) a 3D finite element (FE) model of the BioTac, which complements the experimental dataset with high-density, distributed contact data, 3) neural-network-based mappings from raw BioTac signals to not only low-dimensional experimental data, but also high-density FE deformation fields, and 4) mappings from the FE deformation fields to the raw signals themselves. The high-density data streams can provide a far greater quantity of interpretable information for grasping and manipulation algorithms than previously accessible.

Mandy Xie, Karl Van Wyk, Anqi Li et al.

This paper describes the pragmatic design and construction of geometric fabrics for shaping a robot's task-independent nominal behavior, capturing behavioral components such as obstacle avoidance, joint limit avoidance, redundancy resolution, global navigation heuristics, etc. Geometric fabrics constitute the most concrete incarnation of a new mathematical formulation for reactive behavior called optimization fabrics. Fabrics generalize recent work on Riemannian Motion Policies (RMPs); they add provable stability guarantees and improve design consistency while promoting the intuitive acceleration-based principles of modular design that make RMPs successful. We describe a suite of mathematical modeling tools that practitioners can employ in practice and demonstrate both how to mitigate system complexity by constructing behaviors layer-wise and how to employ these tools to design robust, strongly-generalizing, policies that solve practical problems one would expect to find in industry applications. Our system exhibits intelligent global navigation behaviors expressed entirely as provably stable fabrics with zero planning or state machine governance.

Nathan D. Ratliff, Karl Van Wyk, Mandy Xie et al.

A common approach to the provably stable design of reactive behavior, exemplified by operational space control, is to reduce the problem to the design of virtual classical mechanical systems (energy shaping). This framework is widely used, and through it we gain stability, but at the price of expressivity. This work presents a comprehensive theoretical framework expanding this approach showing that there is a much larger class of differential equations generalizing classical mechanical systems (and the broader class of Lagrangian systems) and greatly expanding their expressivity while maintaining the same governing stability principles. At the core of our framework is a class of differential equations we call fabrics which constitute a behavioral medium across which we can optimize a potential function. These fabrics shape the system's behavior during optimization but still always provably converge to a local minimum, making them a building block of stable behavioral design. We build the theoretical foundations of our framework here and provide a simple empirical demonstration of a practical class of geometric fabrics, which additionally exhibit a natural geometric path consistency making them convenient for flexible and intuitive behavioral design.

Nathan D. Ratliff, Karl Van Wyk, Mandy Xie et al.

Robotics research has found numerous important applications of Riemannian geometry. Despite that, the concept remain challenging to many roboticists because the background material is complex and strikingly foreign. Beyond {\em Riemannian} geometry, there are many natural generalizations in the mathematical literature -- areas such as Finsler geometry and spray geometry -- but those generalizations are largely inaccessible, and as a result there remain few applications within robotics. This paper presents a re-derivation of spray and Finsler geometries we found critical for the development of our recent work on a powerful behavioral design tool we call geometric fabrics. These derivations build from basic tools in advanced calculus and the calculus of variations making them more accessible to a robotics audience than standard presentations. We focus on the pragmatic and calculable results, avoiding the use of tensor notation to appeal to a broader audience, emphasizing geometric path consistency over ideas around connections and curvature. We hope that these derivations will contribute to an increased understanding of generalized nonlinear, and even classical Riemannian, geometry within the robotics community and inspire future research into new applications.

Nathan D. Ratliff, Karl Van Wyk, Mandy Xie et al.

This paper presents a theory of optimization fabrics, second-order differential equations that encode nominal behaviors on a space and can be used to define the behavior of a smooth optimizer. Optimization fabrics can encode commonalities among optimization problems that reflect the structure of the space itself, enabling smooth optimization processes to intelligently navigate each problem even when optimizing simple naive potential functions. Importantly, optimization over a fabric is inherently asymptotically stable. The majority of this paper is dedicated to the development of a tool set for the design and use of a broad class of fabrics called geometric fabrics. Geometric fabrics encode behavior as general nonlinear geometries which are covariant second-order differential equations with a special homogeneity property that ensures their behavior is independent of the system's speed through the medium. A class of Finsler Lagrangian energies can be used to both define how these nonlinear geometries combine with one another and how they react when potential functions force them from their nominal paths. Furthermore, these geometric fabrics are closed under the standard operations of pullback and combination on a transform tree. For behavior representation, this class of geometric fabrics constitutes a broad class of spectral semi-sprays (specs), also known as Riemannian Motion Policies (RMPs) in the context of robotic motion generation, that captures both the intuitive separation between acceleration policy and priority metric critical for modular design and are inherently stable. Therefore, geometric fabrics are safe and easier to use by less experienced behavioral designers. Application of this theory to policy representation and generalization in learning are discussed as well.

Yashraj S. Narang, Karl Van Wyk, Arsalan Mousavian et al.

High-density afferents in the human hand have long been regarded as essential for human grasping and manipulation abilities. In contrast, robotic tactile sensors are typically used to provide low-density contact data, such as center-of-pressure and resultant force. Although useful, this data does not exploit the rich information content that some tactile sensors (e.g., the SynTouch BioTac) naturally provide. This research extends robotic tactile sensing beyond reduced-order models through 1) the automated creation of a precise tactile dataset for the BioTac over diverse physical interactions, 2) a 3D finite element (FE) model of the BioTac, which complements the experimental dataset with high-resolution, distributed contact data, and 3) neural-network-based mappings from raw BioTac signals to low-dimensional experimental data, and more importantly, high-density FE deformation fields. These data streams can provide a far greater quantity of interpretable information for grasping and manipulation algorithms than previously accessible.

Jacky Liang, Ankur Handa, Karl Van Wyk et al.

Tracking the pose of an object while it is being held and manipulated by a robot hand is difficult for vision-based methods due to significant occlusions. Prior works have explored using contact feedback and particle filters to localize in-hand objects. However, they have mostly focused on the static grasp setting and not when the object is in motion, as doing so requires modeling of complex contact dynamics. In this work, we propose using GPU-accelerated parallel robot simulations and derivative-free, sample-based optimizers to track in-hand object poses with contact feedback during manipulation. We use physics simulation as the forward model for robot-object interactions, and the algorithm jointly optimizes for the state and the parameters of the simulations, so they better match with those of the real world. Our method runs in real-time (30Hz) on a single GPU, and it achieves an average point cloud distance error of 6mm in simulation experiments and 13mm in the real-world ones. View experiment videos at https://sites.google.com/view/in-hand-object-pose-tracking/

Ankur Handa, Karl Van Wyk, Wei Yang et al.

Teleoperation offers the possibility of imparting robotic systems with sophisticated reasoning skills, intuition, and creativity to perform tasks. However, current teleoperation solutions for high degree-of-actuation (DoA), multi-fingered robots are generally cost-prohibitive, while low-cost offerings usually provide reduced degrees of control. Herein, a low-cost, vision based teleoperation system, DexPilot, was developed that allows for complete control over the full 23 DoA robotic system by merely observing the bare human hand. DexPilot enables operators to carry out a variety of complex manipulation tasks that go beyond simple pick-and-place operations. This allows for collection of high dimensional, multi-modality, state-action data that can be leveraged in the future to learn sensorimotor policies for challenging manipulation tasks. The system performance was measured through speed and reliability metrics across two human demonstrators on a variety of tasks. The videos of the experiments can be found at https://sites.google.com/view/dex-pilot.

Karl Van Wyk, Joe Falco, Geraldine Cheok

The advancement of simulation-assisted robot programming, automation of high-tolerance assembly operations, and improvement of real-world performance engender a need for positionally accurate robots. Despite tight machining tolerances, good mechanical design, and careful assembly, robotic arms typically exhibit average Cartesian positioning errors of several millimeters. Fortunately, the vast majority of this error can be removed in software by proper calibration of the so-called "zero-offsets" of a robot's joints. This research developed an automated, inexpensive, highly portable, in situ calibration method that fine tunes these kinematic parameters, thereby, improving a robot's average positioning accuracy four-fold throughout its workspace. In particular, a prospective low-cost motion capture system and a benchmark laser tracker were used as reference sensors for robot calibration. Bayesian inference produced optimized zero-offset parameters alongside their uncertainty for data from both reference sensors. Relative and absolute accuracy metrics were proposed and applied for quantifying robot positioning accuracy. Uncertainty analysis of a validated, probabilistic robot model quantified the absolute positioning accuracy throughout its entire workspace. Altogether, three measures of accuracy conclusively revealed multi-fold improvement in the positioning accuracy of the robotic arm. Bayesian inference on motion capture data yielded zero-offsets and accuracy calculations comparable to those derived from laser tracker data, ultimately proving this method's viability towards robot calibration.

Karl Van Wyk, Joe Falco

The existence of tactile afferents sensitive to slip-related mechanical transients in the human hand augments the robustness of grasping through secondary force modulation protocols. Despite this knowledge and the fact that tactile-based slip detection has been researched for decades, robust slip detection is still not an out-of-the-box capability for any commercially available tactile sensor. This research seeks to bridge this gap with a comprehensive study addressing several aspects of slip detection. Key developments include a systematic data collection process yielding millions of sensory data points, the generalized conversion of multivariate-to-univariate sensor output, an insightful spectral analysis of the univariate sensor outputs, and the application of Long Short-Term Memory (LSTM) neural networks on the univariate signals to produce robust slip detectors from three commercially available sensors capable of tactile sensing. The sensing elements underlying these sensors vary in quantity, spatial arrangement, and mechanics, leveraging principles in electro-mechanical resistance, optics, and hydro-acoustics. Critically, slip detection performance of the tactile technologies is quantified through a measurement methodology that unveils the effects of data window size, sampling rate, material type, slip speed, and sensor manufacturing variability. Results indicate that the investigated commercial tactile sensors are inherently capable of high-quality slip detection.

Stefano Carpin, Shuo Liu, Joe Falco et al.

This technical report presents an introduction to different aspects of multi-fingered robot grasping. After having introduced relevant mathematical background for modeling, form and force closure are discussed. Next, we present an overview of various grasp planning algorithms with the objective of illustrating different approaches to solve this problem. Finally, we discuss grasp performance benchmarking.