Uncertainty-Aware Self-Supervised Target-Mass Grasping of Granular Foods

2021/05/27 by Kuniyuki Takahashi, Wilson Ko, Takahashi, Kuniyuki +5
Engineering · #FOS: Computer and information sciences #Modular Robots and Swarm Intelligence #Robot Manipulation and Learning #Robotics (cs.RO) #Soft Robotics and Applications

paper · pdf · doi:10.48550/arxiv.2105.12946

openalex publication_date 2021/05/27 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Food packing industry workers typically pick a target amount of food by hand from a food tray and place them in containers. Since menus are diverse and change frequently, robots must adapt and learn to handle new foods in a short time-span. Learning to grasp a specific amount of granular food requires a large training dataset, which is challenging to collect reasonably quickly. In this study, we propose ways to reduce the necessary amount of training data by augmenting a deep neural network with models that estimate its uncertainty through self-supervised learning. To further reduce human effort, we devise a data collection system that automatically generates labels. We build on the idea that we can grasp sufficiently well if there is at least one low-uncertainty (high-confidence) grasp point among the various grasp point candidates. We evaluate the methods we propose in this work on a variety of granular foods -- coffee beans, rice, oatmeal and peanuts -- each of which has a different size, shape and material properties such as volumetric mass density or friction. For these foods, we show significantly improved grasp accuracy of user-specified target masses using smaller datasets by incorporating uncertainty.

Citations

• Bayesian Convolutional Neural Networks with Bernoulli Approximate Variational Inference
• An overview of 3D object grasp synthesis algorithms
• Dropout as a Bayesian Approximation: Appendix
• What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
• Exploration by Random Network Distillation
• Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics
• Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles

Related

• Playing with Food: Learning Food Item Representations through\n Interactive Exploration
• Cluttered Food Grasping with Adaptive Fingers and Synthetic-Data Trained Object Detection
• Adaptive Robot-Assisted Feeding: An Online Learning Framework for Acquiring Previously Unseen Food Items
• GRITS: A Spillage-Aware Guided Diffusion Policy for Robot Food Scooping Tasks
• Learning Bimanual Scooping Policies for Food Acquisition