Dynamic Object Manipulation Through Active Vision: a Reinforcement Learning Approach

EasyChair Preprint 11977

Abstract

In the realm of robotics and artificial intelligence, the manipulation of objects in dynamic and cluttered environments poses significant challenges. Traditional approaches often struggle to adapt to changing scenarios, particularly when objects become occluded or undergo unpredictable movements. This paper presents a novel approach to address these challenges by integrating reinforcement learning with active vision techniques for dynamic object manipulation tasks. The proposed framework leverages the synergy between reinforcement learning algorithms and active vision strategies to enable robots to autonomously learn and adapt their manipulation behaviors in real-time.

Keyphrases: Robotic Manipulation, manipulation, occlusions

@booklet{EasyChair:11977,
  author    = {Jane Smith and Chen Liu},
  title     = {Dynamic Object Manipulation Through Active Vision: a Reinforcement Learning Approach},
  howpublished = {EasyChair Preprint 11977},
  year      = {EasyChair, 2024}}