Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Recent Advances of Deep Robotic Affordance Learning: A Reinforcement Learning Perspective

Xintong Yang, Ze Ji, Jing Wu, Yu-kun Lai

TL;DR

A short review of the recent developments of deep robotic affordance learning (DRAL), which aims to develop data-driven methods that use the concept of affordance to aid in robotic tasks.

Abstract

As a popular concept proposed in the field of psychology, affordance has been regarded as one of the important abilities that enable humans to understand and interact with the environment. Briefly, it captures the possibilities and effects of the actions of an agent applied to a specific object or, more generally, a part of the environment. This paper provides a short review of the recent developments of deep robotic affordance learning (DRAL), which aims to develop data-driven methods that use the concept of affordance to aid in robotic tasks. We first classify these papers from a reinforcement learning (RL) perspective, and draw connections between RL and affordances. The technical details of each category are discussed and their limitations identified. We further summarise them and identify future challenges from the aspects of observations, actions, affordance representation, data-collection and real-world deployment. A final remark is given at the end to propose a promising future direction of the RL-based affordance definition to include the predictions of arbitrary action consequences.

Recent Advances of Deep Robotic Affordance Learning: A Reinforcement Learning Perspective

TL;DR

A short review of the recent developments of deep robotic affordance learning (DRAL), which aims to develop data-driven methods that use the concept of affordance to aid in robotic tasks.

Abstract

As a popular concept proposed in the field of psychology, affordance has been regarded as one of the important abilities that enable humans to understand and interact with the environment. Briefly, it captures the possibilities and effects of the actions of an agent applied to a specific object or, more generally, a part of the environment. This paper provides a short review of the recent developments of deep robotic affordance learning (DRAL), which aims to develop data-driven methods that use the concept of affordance to aid in robotic tasks. We first classify these papers from a reinforcement learning (RL) perspective, and draw connections between RL and affordances. The technical details of each category are discussed and their limitations identified. We further summarise them and identify future challenges from the aspects of observations, actions, affordance representation, data-collection and real-world deployment. A final remark is given at the end to propose a promising future direction of the RL-based affordance definition to include the predictions of arbitrary action consequences.
Paper Structure (16 sections, 2 equations, 4 figures, 2 tables)

This paper contains 16 sections, 2 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Affordance definition in MDPs
  3. Modelling Action Possibilities
  4. Image segmentation
  5. Action scores
  6. Summary and limitations
  7. Keypoint Affordance
  8. Modelling Action Possibilities and Effects
  9. Discussions and challenges
  10. Observation
  11. Action
  12. Affordance Representations
  13. Data collection
  14. Real-world deployment
  15. Conclusion
  16. ...and 1 more sections

Figures (4)

  • Figure 1: Segmented image from chu2019toward. Red parts afford grasping, orange afford supporting, deep blue afford containing, blue afford wrap-grasping, and purple afford pounding.
  • Figure 2: Examples of action score prediction.
  • Figure 3: Category-level keypoint detection from manuelli2019kpam. (a) Detected keypoints for different cups in planning; (b) keypoint detection; (c) grasping; (d) hanging.
  • Figure 4: The multi-step tool-use task designed to evaluate the Deep Affordance Foresight method proposed in xu2020daf. The robot needs to decide which end of the L-shape stick to grasp for reaching the red block or push the blue block out of the tube.