UC Berkeley AUTOLab

UC Berkeley's AUTOLAB, directed by Professor Ken Goldberg, is a leading center for research in robotics and automation sciences. The lab supports approximately 30 postdocs, PhD students, and undergrads pursuing projects in Robust Robot Grasping and Manipulation for Warehouses, Homes, and Robot-Assisted Surgery. We focus on methods at the intersection of analytic theory (eg, geometric and statistical models, motion planning) and deep learning. AUTOLAB Research Papers: http://goldberg.berkeley.edu/pubs/ Virtual-Lab-Tour: https://drive.google.com/open?id=17o7RpbO7eb0O772fITOQOnImCUv5BNVa Our research is sponsored in part by NSF, NDSEG, USDA, Google, Siemens, HP, Honda, Autodesk, Toyota Research Institute, Amazon Robotics, Intel, Samsung, Cisco, and Intuitive Surgical. The AUTOLAB is located in 2111 Etcheverry Hall and 8015 Berkeley Way West, UC Berkeley. A Sample of Ongoing Projects: Dex-Net for Robust Robot Grasping. Combining analytic mechanics theory with Deep Learning using sim2real at scale facilitates highly efficient, robust, grasping that generalizes surprisingly well to new objects. Cloud Robotics. We are exploring how the ability to access remote data, code, and processing can facilitate scalable robot learning. Safety Augmented Value Estimation from Demonstrations (SAVED): Safe Deep Model-Based RL for Sparse Cost Robotic Tasks. Addressing the difficulty in hand-engineering a dense cost function, which can lead to unintended behavior, and dynamical uncertainty, which makes exploration and constraint satisfaction challenging. Deep Learning from Demonstrations. We are developing new methods where humans can efficiently teach robots to robustly perform tasks such as surgical needle insertion, grasping in clutter, part singulation, rope-tying and assembly. Robot-Assisted Surgery. We're developing new geometric models and algorithms for automating surgical subtasks such as suturing and debridement using the da Vinci surgical robot.