A5050342525- Robot Manipulation and Learning
- Soft Robotics and Applications
- Reinforcement Learning in Robotics
- Robotic Path Planning Algorithms
- Modular Robots and Swarm Intelligence
- Robotics and Sensor-Based Localization
- Surgical Simulation and Training
- Robotics and Automated Systems
- Robotic Mechanisms and Dynamics
- 3D Shape Modeling and Analysis
- Human Pose and Action Recognition
- Anatomy and Medical Technology
- Advanced Bandit Algorithms Research
- Augmented Reality Applications
- Multimodal Machine Learning Applications
- IoT and Edge/Fog Computing
- Hand Gesture Recognition Systems
- Adversarial Robustness in Machine Learning
- Smart Agriculture and AI
- Muscle activation and electromyography studies
- Advanced Vision and Imaging
- Manufacturing Process and Optimization
- Machine Learning and Algorithms
- Teleoperation and Haptic Systems
- Advanced Neural Network Applications
University of California, Berkeley
2015-2024
Berkeley Systems (United States)
2017-2024
Berkeley College
2015-2024
Stanford University
2021
Scuola Superiore Sant'Anna
2021
Max Planck Institute for Intelligent Systems
2021
Fischer (Germany)
2021
University of Stuttgart
2021
Santa Clara University
2020-2021
Shanghai Jiao Tong University
2021
To reduce data collection time for deep learning of robust robotic grasp plans, we explore training from a synthetic dataset 6.7 million point clouds, grasps, and analytic metrics generated thousands 3D models Dex-Net 1.0 in randomized poses on table.We use the resulting dataset, 2.0, to train Grasp Quality Convolutional Neural Network (GQ-CNN) model that rapidly predicts probability success grasps depth images, where are specified as planar position, angle, gripper relative an RGB-D...
COVID-19 may drive sustained research in robotics to address risks of infectious diseases.
Universal picking (UP), or reliable robot grasping of a diverse range novel objects from heaps, is grand challenge for e-commerce order fulfillment, manufacturing, inspection, and home service robots. Optimizing the rate, reliability, UP difficult due to inherent uncertainty in sensing, control, contact physics. This paper explores "ambidextrous" grasping, where two more heterogeneous grippers are used. We present Dexterity Network (Dex-Net) 4.0, substantial extension previous versions...
Imitation learning is a powerful paradigm for robot skill acquisition. However, obtaining demonstrations suitable policy that maps from raw pixels to actions can be challenging. In this paper we describe how consumer-grade Virtual Reality headsets and hand tracking hardware used naturally teleoperate robots perform complex tasks. We also imitation learn deep neural network policies (mapping actions) acquire the demonstrated skills. Our experiments showcase effectiveness of our approach visuomotor
This paper presents the Dexterity Network (Dex-Net) 1.0, a dataset of 3D object models and sampling-based planning algorithm to explore how Cloud Robotics can be used for robust grasp planning. The uses Multi- Armed Bandit model with correlated rewards leverage prior grasps in growing that currently includes over 10,000 unique 2.5 million parallel-jaw grasps. Each an estimate probability force closure under uncertainty gripper pose friction. Dex-Net 1.0 Multi-View Convolutional Neural...
Vacuum-based end effectors are widely used in industry and often preferred over parallel-jaw multifinger grippers due to their ability lift objects with a single point of contact. Suction grasp planners target planar surfaces on clouds near the estimated centroid an object. In this paper, we propose compliant suction contact model that computes quality seal between cup local surface measure resist external gravity wrench. To characterize grasps, estimate robustness perturbations end-effector...
To reduce data collection time for deep learning of robust robotic grasp plans, we explore training from a synthetic dataset 6.7 million point clouds, grasps, and analytic metrics generated thousands 3D models Dex-Net 1.0 in randomized poses on table. We use the resulting dataset, 2.0, to train Grasp Quality Convolutional Neural Network (GQ-CNN) model that rapidly predicts probability success grasps depth images, where are specified as planar position, angle, gripper relative an RGB-D...
Reinforcement learning (RL) algorithms involve the deep nesting of highly irregular computation patterns, each which typically exhibits opportunities for distributed computation. We argue distributing RL components in a composable way by adapting top-down hierarchical control, thereby encapsulating parallelism and resource requirements within short-running compute tasks. demonstrate benefits this principle through RLlib: library that provides scalable software primitives RL. These enable...
Analysts often clean dirty data iteratively--cleaning some data, executing the analysis, and then cleaning more based on results. We explore iterative process in context of statistical model training, which is an increasingly popular form analytics. propose ActiveClean, allows for progressive modeling problems while preserving convergence guarantees. ActiveClean supports important class models called convex loss (e.g., linear regression SVMs), prioritizes those records likely to affect...
With the increasing commoditization of computer vision, speech recognition and machine translation systems widespread deployment learning-based back-end technologies such as digital advertising intelligent infrastructures, AI (Artificial Intelligence) has moved from research labs to production. These changes have been made possible by unprecedented levels data computation, methodological advances in learning, innovations software architectures, broad accessibility these technologies. The...
The ability to segment unknown objects in depth images has potential enhance robot skills grasping and object tracking. Recent computer vision research demonstrated that Mask R-CNN can be trained specific categories of RGB when massive hand-labeled datasets are available. As generating these is time-consuming, we instead train with synthetic images. Many robots now use sensors, recent results suggest training on data transfer successfully the real world. We present a method for automated...
Safety remains a central obstacle preventing widespread use of RL in the real world: learning new tasks uncertain environments requires extensive exploration, but safety limiting exploration. We propose Recovery RL, an algorithm which navigates this tradeoff by (1) leveraging offline data to learn about constraint violating zones <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">before</i> policy and (2)...
Humans describe the physical world using natural language to refer specific 3D locations based on a vast range of properties: visual appearance, semantics, abstract associations, or actionable affordances. In this work we propose Language Embedded Radiance Fields (LERFs), method for grounding embeddings from off-the-shelf models like CLIP into NeRF, which enable these types open-ended queries in 3D. LERF learns dense, multi-scale field inside NeRF by volume rendering along training rays,...
Artificial intelligence (AI) applications in medical robots are bringing a new era to medicine. Advanced can perform diagnostic and surgical procedures, aid rehabilitation, provide symbiotic prosthetics replace limbs. The technology used these devices, including computer vision, image analysis, haptics, navigation, precise manipulation, machine learning (ML) , could allow autonomous carry out imaging, remote surgery, subtasks, or even entire procedures. Moreover, AI rehabilitation devices...
Case reports and cohort studies have linked bisphosphonate therapy osteonecrosis of the jaws (ONJ), but neither causality nor specific risks for lesion development been clearly established. We conducted a 1:3 case-control study with three dental Practice-based Research Networks, using dentist questionnaires patient interviews collection data on therapy, demographics, co-morbidities, medical treatments. Multivariable logistic regression analyses tested associations between use other risk...
For a responsive audio art installation in skylit atrium, we introduce single-camera statistical segmentation and tracking algorithm. The algorithm combines background image estimation, per-pixel Bayesian segmentation, an approximate solution to the multi-target problem using bank of Kalman filters Gale-Shapley matching. A heuristic confidence model enables selective filtering tracks based on dynamic data. We demonstrate that our has improved recall F <sub...
For supervised automation of multi-throw suturing in Robot-Assisted Minimally Invasive Surgery, we present a novel mechanical needle guide and framework for optimizing size, trajectory, control parameters using sequential convex programming. The Suture Needle Angular Positioner (SNAP) results 3x error reduction the pose estimate comparison with standard actuator. We evaluate algorithm SNAP on da Vinci Research Kit tissue phantoms compare completion time that humans from JIGSAWS dataset [5]....
In the Fundamentals of Laparoscopic Surgery (FLS) standard medical training regimen, Pattern Cutting task requires residents to demonstrate proficiency by maneuvering two tools, surgical scissors and tissue gripper, accurately cut a circular pattern on gauze suspended at corners. Accuracy cutting depends tensioning, wherein gripper pinches point in R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> pulls induce maintain tension material...
Rapid and reliable robot grasping for a diverse set of objects has applications from warehouse automation to home decluttering. One promising approach is learn deep policies synthetic training datasets point clouds, grasps, rewards sampled using analytic models with stochastic noise domain randomization. In this letter, we explore how the distribution examples affects rate reliability learned policy. We propose data sampling that combines grasps policy action guiding samples robust...
Rearranging and manipulating deformable objects such as cables, fabrics, bags is a long-standing challenge in robotic manipulation. The complex dynamics high-dimensional configuration spaces of deformables, compared to rigid objects, make manipulation difficult not only for multi-step planning, but even goal specification. Goals cannot be easily specified object poses, may involve relative spatial relations "place the item inside bag". In this work, we develop suite simulated benchmarks with...
Background: Growing evidence shows that fixed, nonpersonalized daily step goals can discourage individuals, resulting in unchanged or even reduced physical activity. Objective: The aim of this randomized controlled trial (RCT) was to evaluate the efficacy an automated mobile phone–based personalized and adaptive goal-setting intervention using machine learning as compared with active control steady 10,000. Methods: In 10-week RCT, 64 participants were recruited via email announcements...
Robotic manipulation of deformable 1D objects such as ropes, cables, and hoses is challenging due to the lack high-fidelity analytic models large configuration spaces. Furthermore, learning end-to-end policies directly from images physical interaction requires significant time on a robot can fail generalize across tasks. We address these challenges using interpretable deep visual representations for rope, extending recent work dense object descriptors manipulation. This facilitates design...