Eugen Solowjow

Brian Zhu, Philipp Schmitt, Philine Meister et al.

Vision-Language-Action (VLA) policies have shown promising manipulation capabilities, yet their practical impact is often limited by the reliability demands of real-world deployment. We present a deployment study of an industrial packaging task at Siemens Factory (GWE, Erlangen, Germany), where a robot must pick a transparent accessory bag from a cluttered pile, insert it into the remaining cavity of a cardboard package, and ensure that the bag and its contents remain below the closing plane. Our goal is to understand the practical effort required to adapt a pretrained Pi0.5 policy to a single factory-floor task through iterative fine-tuning and deployment-driven refinement. The pipeline consists of repeated loops of data collection, curation, fine-tuning, evaluation, and targeted recovery data collection. We have accumulated 2535 episodes (10 hours) from the on-site factory settings. In this paper, we contribute an empirical account of a factory-floor VLA deployment, highlighting recurring failure modes and lessons that inform how to improve the deployment workflow.

Ashvin Nair, Brian Zhu, Gokul Narayanan et al.

Learning-based methods in robotics hold the promise of generalization, but what can be done if a learned policy does not generalize to a new situation? In principle, if an agent can at least evaluate its own success (i.e., with a reward classifier that generalizes well even when the policy does not), it could actively practice the task and finetune the policy in this situation. We study this problem in the setting of industrial insertion tasks, such as inserting connectors in sockets and setting screws. Existing algorithms rely on precise localization of the connector or socket and carefully managed physical setups, such as assembly lines, to succeed at the task. But in unstructured environments such as homes or even some industrial settings, robots cannot rely on precise localization and may be tasked with previously unseen connectors. Offline reinforcement learning on a variety of connector insertion tasks is a potential solution, but what if the robot is tasked with inserting previously unseen connector? In such a scenario, we will still need methods that can robustly solve such tasks with online practice. One of the main observations we make in this work is that, with a suitable representation learning and domain generalization approach, it can be significantly easier for the reward function to generalize to a new but structurally similar task (e.g., inserting a new type of connector) than for the policy. This means that a learned reward function can be used to facilitate the finetuning of the robot's policy in situations where the policy fails to generalize in zero shot, but the reward function generalizes successfully. We show that such an approach can be instantiated in the real world, pretrained on 50 different connectors, and successfully finetuned to new connectors via the learned reward function. Videos can be viewed at https://sites.google.com/view/learningonthejob

Wisdom C. Agboh, Satvik Sharma, Kishore Srinivas et al.

We consider a decluttering problem where multiple rigid convex polygonal objects rest in randomly placed positions and orientations on a planar surface and must be efficiently transported to a packing box using both single and multi-object grasps. Prior work considered frictionless multi-object grasping. In this paper, we introduce friction to increase the number of potential grasps for a given group of objects, and thus increase picks per hour. We train a neural network using real examples to plan robust multi-object grasps. In physical experiments, we find a 13.7% increase in success rate, a 1.6x increase in picks per hour, and a 6.3x decrease in grasp planning time compared to prior work on multi-object grasping. Compared to single-object grasping, we find a 3.1x increase in picks per hour.

Gaurav Datta, Ryan Hoque, Anrui Gu et al.

Imitation learning has been applied to a range of robotic tasks, but can struggle when robots encounter edge cases that are not represented in the training data (i.e., distribution shift). Interactive fleet learning (IFL) mitigates distribution shift by allowing robots to access remote human supervisors during task execution and learn from them over time, but different supervisors may demonstrate the task in different ways. Recent work proposes Implicit Behavior Cloning (IBC), which is able to represent multimodal demonstrations using energy-based models (EBMs). In this work, we propose Implicit Interactive Fleet Learning (IIFL), an algorithm that builds on IBC for interactive imitation learning from multiple heterogeneous human supervisors. A key insight in IIFL is a novel approach for uncertainty quantification in EBMs using Jeffreys divergence. While IIFL is more computationally expensive than explicit methods, results suggest that IIFL achieves a 2.8x higher success rate in simulation experiments and a 4.5x higher return on human effort in a physical block pushing task over (Explicit) IFL, IBC, and other baselines.

Gokul Narayanan, Yash Shahapurkar, Melih Erdogan et al.

Industrial automation is at a pivotal moment, as Physical AI is driving a transition from rigid, hand-engineered automation systems toward more flexible and adaptive systems. This shift has created a growing demand for large-scale, real-world robot demonstration data, making teleoperation an increasingly important mechanism for data collection. However, high-quality teleoperated demonstrations remain difficult to obtain in practice, as novice operators often produce episodes that are task-successful but suboptimal for downstream use due to inefficient motion, repeated corrections, or operation near robot joint limits. We present a Data Quality Assessment and Feedback (DQAF) framework that closes the loop in teleoperation by providing immediate post-episode feedback grounded in semantic task progress and robot telemetry. The framework extracts quality relevant signals such as sub-task progress, motion smoothness, stalls, kinematic limits and converts them into structured quality assessments and actionable natural-language feedback. Unlike binary success or failure feedback, the proposed system explains why an episode is suboptimal and highlights specific behaviors to correct in the next trial. We evaluate the framework through a diagnostic validation study and a pilot user study. In the validation study, the system is compared with a human reviewer during dataset curation, producing rejection reasons and actionable feedback for improvement. In the pilot study with three novice operators across two manipulation tasks, the operator who received the systems immediate, automated post-episode feedback improved faster than those who did not, producing higher-quality demonstrations sooner.

Andrew Benton, Eugen Solowjow, Prithvi Akella

A robotic behavior model that can reliably generate behaviors from natural language inputs in real time would substantially expedite the adoption of industrial robots due to enhanced system flexibility. To facilitate these efforts, we construct a framework in which learned behaviors, created by a natural language abstractor, are verifiable by construction. Leveraging recent advancements in motion primitives and probabilistic verification, we construct a natural-language behavior abstractor that generates behaviors by synthesizing a directed graph over the provided motion primitives. If these component motion primitives are constructed according to the criteria we specify, the resulting behaviors are probabilistically verifiable. We demonstrate this verifiable behavior generation capacity in both simulation on an exploration task and on hardware with a robot scooping granular media.

Shivin Devgon, Jeffrey Ichnowski, Michael Danielczuk et al.

In industrial part kitting, 3D objects are inserted into cavities for transportation or subsequent assembly. Kitting is a critical step as it can decrease downstream processing and handling times and enable lower storage and shipping costs. We present Kit-Net, a framework for kitting previously unseen 3D objects into cavities given depth images of both the target cavity and an object held by a gripper in an unknown initial orientation. Kit-Net uses self-supervised deep learning and data augmentation to train a convolutional neural network (CNN) to robustly estimate 3D rotations between objects and matching concave or convex cavities using a large training dataset of simulated depth images pairs. Kit-Net then uses the trained CNN to implement a controller to orient and position novel objects for insertion into novel prismatic and conformal 3D cavities. Experiments in simulation suggest that Kit-Net can orient objects to have a 98.9% average intersection volume between the object mesh and that of the target cavity. Physical experiments with industrial objects succeed in 18% of trials using a baseline method and in 63% of trials with Kit-Net. Video, code, and data are available at https://github.com/BerkeleyAutomation/Kit-Net.

Jianlan Luo, Eugen Solowjow, Chengtao Wen et al.

Precise robotic manipulation skills are desirable in many industrial settings, reinforcement learning (RL) methods hold the promise of acquiring these skills autonomously. In this paper, we explicitly consider incorporating operational space force/torque information into reinforcement learning; this is motivated by humans heuristically mapping perceived forces to control actions, which results in completing high-precision tasks in a fairly easy manner. Our approach combines RL with force/torque information by incorporating a proper operational space force controller; where we also exploit different ablations on processing this information. Moreover, we propose a neural network architecture that generalizes to reasonable variations of the environment. We evaluate our method on the open-source Siemens Robot Learning Challenge, which requires precise and delicate force-controlled behavior to assemble a tight-fit gear wheel set.

Simeon Adebola, Chung Min Kim, Justin Kerr et al.

Commercial plant phenotyping systems using fixed cameras cannot perceive many plant details due to leaf occlusion. In this paper, we present Botany-Bot, a system for building detailed "annotated digital twins" of living plants using two stereo cameras, a digital turntable inside a lightbox, an industrial robot arm, and 3D segmentated Gaussian Splat models. We also present robot algorithms for manipulating leaves to take high-resolution indexable images of occluded details such as stem buds and the underside/topside of leaves. Results from experiments suggest that Botany-Bot can segment leaves with 90.8% accuracy, detect leaves with 86.2% accuracy, lift/push leaves with 77.9% accuracy, and take detailed overside/underside images with 77.3% accuracy. Code, videos, and datasets are available at https://berkeleyautomation.github.io/Botany-Bot/.

Letian Fu, Michael Danielczuk, Ashwin Balakrishna et al.

While deep learning has enabled significant progress in designing general purpose robot grasping systems, there remain objects which still pose challenges for these systems. Recent work on Exploratory Grasping has formalized the problem of systematically exploring grasps on these adversarial objects and explored a multi-armed bandit model for identifying high-quality grasps on each object stable pose. However, these systems are still limited to exploring a small number or grasps on each object. We present Learned Efficient Grasp Sets (LEGS), an algorithm that efficiently explores thousands of possible grasps by maintaining small active sets of promising grasps and determining when it can stop exploring the object with high confidence. Experiments suggest that LEGS can identify a high-quality grasp more efficiently than prior algorithms which do not use active sets. In simulation experiments, we measure the gap between the success probability of the best grasp identified by LEGS, baselines, and the most-robust grasp (verified ground truth). After 3000 exploration steps, LEGS outperforms baseline algorithms on 10/14 and 25/39 objects on the Dex-Net Adversarial and EGAD! datasets respectively. We then evaluate LEGS in physical experiments; trials on 3 challenging objects suggest that LEGS converges to high-performing grasps significantly faster than baselines. See https://sites.google.com/view/legs-exp-grasping for supplemental material and videos.

Martin A. Sehr, Wei Xi Xia, Prithvi Akella et al.

Active perception systems maximizing information gain to support both monitoring and decision making have seen considerable application in recent work. In this paper, we propose and demonstrate a method of acquiring and extrapolating information in an active sensory system through use of a Bayesian Filter. Our approach is motivated by manufacturing processes, where automated visual tracking of system states may aid in fault diagnosis, certification of parts and safety; in extreme cases, our approach may enable novel manufacturing processes relying on monitoring solutions beyond passive perception. We demonstrate how using a Bayesian Filter in active perception scenarios permits reasoning about future actions based on measured as well as unmeasured but propagated state elements, thereby increasing substantially the quality of information available to decision making algorithms used in control of overarching processes. We demonstrate use of our active perception system in physical experiments, where we use a time-varying Kalman Filter to resolve uncertainty for a representative system capturing in additive manufacturing.

Gerrit Schoettler, Ashvin Nair, Juan Aparicio Ojea et al.

Robotic insertion tasks are characterized by contact and friction mechanics, making them challenging for conventional feedback control methods due to unmodeled physical effects. Reinforcement learning (RL) is a promising approach for learning control policies in such settings. However, RL can be unsafe during exploration and might require a large amount of real-world training data, which is expensive to collect. In this paper, we study how to use meta-reinforcement learning to solve the bulk of the problem in simulation by solving a family of simulated industrial insertion tasks and then adapt policies quickly in the real world. We demonstrate our approach by training an agent to successfully perform challenging real-world insertion tasks using less than 20 trials of real-world experience. Videos and other material are available at https://pearl-insertion.github.io/

Eugen Solowjow, Ines Ugalde, Yash Shahapurkar et al.

Universal grasping of a diverse range of previously unseen objects from heaps is a grand challenge in e-commerce order fulfillment, manufacturing, and home service robotics. Recently, deep learning based grasping approaches have demonstrated results that make them increasingly interesting for industrial deployments. This paper explores the problem from an automation systems point-of-view. We develop a robotics grasping system using Dex-Net, which is fully integrated at the controller level. Two neural networks are deployed on a novel industrial AI hardware acceleration module close to a PLC with a power footprint of less than 10 W for the overall system. The software is tightly integrated with the hardware allowing for fast and efficient data processing and real-time communication. The success rate of grasping an object form a bin is up to 95 percent with more than 350 picks per hour, if object and receptive bins are in close proximity. The system was presented at the Hannover Fair 2019 (world s largest industrial trade fair) and other events, where it performed over 5,000 grasps per event.

Lin Shao, Fabio Ferreira, Mikael Jorda et al.

To achieve a successful grasp, gripper attributes such as its geometry and kinematics play a role as important as the object geometry. The majority of previous work has focused on developing grasp methods that generalize over novel object geometry but are specific to a certain robot hand. We propose UniGrasp, an efficient data-driven grasp synthesis method that considers both the object geometry and gripper attributes as inputs. UniGrasp is based on a novel deep neural network architecture that selects sets of contact points from the input point cloud of the object. The proposed model is trained on a large dataset to produce contact points that are in force closure and reachable by the robot hand. By using contact points as output, we can transfer between a diverse set of multifingered robotic hands. Our model produces over 90% valid contact points in Top10 predictions in simulation and more than 90% successful grasps in real world experiments for various known two-fingered and three-fingered grippers. Our model also achieves 93%, 83% and 90% successful grasps in real world experiments for an unseen two-fingered gripper and two unseen multi-fingered anthropomorphic robotic hands.

Gerrit Schoettler, Ashvin Nair, Jianlan Luo et al.

Connector insertion and many other tasks commonly found in modern manufacturing settings involve complex contact dynamics and friction. Since it is difficult to capture related physical effects with first-order modeling, traditional control methods often result in brittle and inaccurate controllers, which have to be manually tuned. Reinforcement learning (RL) methods have been demonstrated to be capable of learning controllers in such environments from autonomous interaction with the environment, but running RL algorithms in the real world poses sample efficiency and safety challenges. Moreover, in practical real-world settings we cannot assume access to perfect state information or dense reward signals. In this paper, we consider a variety of difficult industrial insertion tasks with visual inputs and different natural reward specifications, namely sparse rewards and goal images. We show that methods that combine RL with prior information, such as classical controllers or demonstrations, can solve these tasks from a reasonable amount of real-world interaction.

Xinyi Ren, Jianlan Luo, Eugen Solowjow et al.

Accurate state estimation is a fundamental component of robotic control. In robotic manipulation tasks, as is our focus in this work, state estimation is essential for identifying the positions of objects in the scene, forming the basis of the manipulation plan. However, pose estimation typically requires expensive 3D cameras or additional instrumentation such as fiducial markers to perform accurately. Recently, Tobin et al.~introduced an approach to pose estimation based on domain randomization, where a neural network is trained to predict pose directly from a 2D image of the scene. The network is trained on computer-generated images with a high variation in textures and lighting, thereby generalizing to real-world images. In this work, we investigate how to improve the accuracy of domain randomization based pose estimation. Our main idea is that active perception -- moving the robot to get a better estimate of pose -- can be trained in simulation and transferred to real using domain randomization. In our approach, the robot trains in a domain-randomized simulation how to estimate pose from a \emph{sequence} of images. We show that our approach can significantly improve the accuracy of standard pose estimation in several scenarios: when the robot holding an object moves, when reference objects are moved in the scene, or when the camera is moved around the object.

Tobias Johannink, Shikhar Bahl, Ashvin Nair et al.

Conventional feedback control methods can solve various types of robot control problems very efficiently by capturing the structure with explicit models, such as rigid body equations of motion. However, many control problems in modern manufacturing deal with contacts and friction, which are difficult to capture with first-order physical modeling. Hence, applying control design methodologies to these kinds of problems often results in brittle and inaccurate controllers, which have to be manually tuned for deployment. Reinforcement learning (RL) methods have been demonstrated to be capable of learning continuous robot controllers from interactions with the environment, even for problems that include friction and contacts. In this paper, we study how we can solve difficult control problems in the real world by decomposing them into a part that is solved efficiently by conventional feedback control methods, and the residual which is solved with RL. The final control policy is a superposition of both control signals. We demonstrate our approach by training an agent to successfully perform a real-world block assembly task involving contacts and unstable objects.