Grasp It Like a Pro 2.0: A Data-Driven Approach Exploiting Basic Shape Decomposition and Human Data for Grasping Unknown Objects

Palleschi, Alessandro; Angelini, Franco; Gabellieri, Chiara; Park, Do Won; Pallottino, Lucia; Bicchi, Antonio; Garabini, Manolo

doi:10.1109/tro.2023.3286115

Cited by 7 publications

(2 citation statements)

References 65 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When the end-effector and the target point are separated by a specific distance, the rewards are modified and intermediate rewards are adjusted. The reward setting is shown in Equation (8).…”

Section: Reward Reshapingmentioning

confidence: 99%

“…This entails developing intricate mathematical models to capture dynamic factors such as sliding friction, gravity, inertia, and motion planning for moving objects. Additionally, utilizing known object shapes, poses, materials, and desired trajectories for computation purposes can be expensive and challenging to adapt to new objects and environments [8].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Towards Multi-Objective Object Push-Grasp Policy Based on Maximum Entropy Deep Reinforcement Learning under Sparse Rewards

Zhang,

2024

Entropy

View full text Add to dashboard Cite

In unstructured environments, robots need to deal with a wide variety of objects with diverse shapes, and often, the instances of these objects are unknown. Traditional methods rely on training with large-scale labeled data, but in environments with continuous and high-dimensional state spaces, the data become sparse, leading to weak generalization ability of the trained models when transferred to real-world applications. To address this challenge, we present an innovative maximum entropy Deep Q-Network (ME-DQN), which leverages an attention mechanism. The framework solves complex and sparse reward tasks through probabilistic reasoning while eliminating the trouble of adjusting hyper-parameters. This approach aims to merge the robust feature extraction capabilities of Fully Convolutional Networks (FCNs) with the efficient feature selection of the attention mechanism across diverse task scenarios. By integrating an advantage function with the reasoning and decision-making of deep reinforcement learning, ME-DQN propels the frontier of robotic grasping and expands the boundaries of intelligent perception and grasping decision-making in unstructured environments. Our simulations demonstrate a remarkable grasping success rate of 91.6%, while maintaining excellent generalization performance in the real world.

show abstract

Section: Reward Reshapingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%