Grasp Planning using Probabilistic Inference - Qingkai Lu

Grasping (deciding how to pick up an object) is a difficult problem in Robotics. In this talk, I will present how we plan grasps using probabilistic inference. I will first talk about our novel approach to multi-fingered grasp planning leveraging learned deep neural network models. We train a convolutional neural network to predict grasp success as a function of the visual information of an object and the grasp configuration. We can then formulate grasp planning as inferring the grasp configuration which maximizes the probability of grasp success. Our experimental results show that our planning method outperforms existing planning methods for neural networks. Then I will introduce our probabilistic grasp planner that explicitly models grasp type for planning high-quality precision and power grasps in real-time. We compare our learned grasp model with a model that does not encode type and show modeling grasp type helps to plan better grasps. At the end of the talk, I will briefly discuss our on-going active learning work for grasping.

Qingkai Lu is a PhD student in School of Computing, University of Utah. His research focuses on robot grasp learning and planning. He has published two top tier research papers in robotics. He did a research internship at Amazon Robotics last summer. In this talk, he will talk about his work on grasp learning and planning.

Location is provided by HireVue
Food will be provided by Google Cloud Platform