Research Analyzer
← Back ICRA 2026

The Developments and Challenges towards Dexterous and Embodied Robotic Manipulation: A Survey

Gaofeng Li, Ruize Wang, Peisen Xu, Qi Ye, Jiming Chen

PDF

AI summary

Key figure (auto-extracted from paper)
Achieving human-like robotic dexterity requires a transition to embodied intelligence that integrates multi-fingered hardware with diverse data collection and skill-learning frameworks.
Dexterous Manipulation Embodied Intelligence Multi-fingered Hands Data Collection Reinforcement Learning Imitation Learning

Problem

Robotic manipulation remains an open problem because high-degree-of-freedom dexterous hands are difficult to control and lack the massive, high-quality datasets needed for effective learning.

Approach

A systematic survey that categorizes the evolution of robotic manipulation into three historical stages and reviews current advances in data collection paradigms and skill-learning frameworks.

Key results

  • Categorized manipulation evolution into Mechanical Programming, Closed-Loop Control, and Embodied Intelligence stages
  • Reviewed data collection methods via simulation platforms, human demonstrations, and teleoperation
  • Analyzed skill-learning frameworks focusing on imitation learning and reinforcement learning
  • Identified three key challenges restricting the development of dexterous robotic manipulation

Why it matters

This survey provides a comprehensive roadmap for researchers to move from simple grippers to human-like dexterity in unstructured environments.

Abstract

Achieving human-like dexterous robotic manipula- tion remains a central goal and a pivotal challenge in robotics. The development of Artificial Intelligence (AI) has allowed rapid progress in robotic manipulation. This survey summarizes the evolution of robotic manipulation from mechanical programming to embodied intelligence, alongside the transition from simple grippers to multi-fingered dexterous hands, outlining key char- acteristics and main challenges. Focusing on the current stage of embodied dexterous manipulation, we highlight recent advances in two critical areas: dexterous manipulation data collection (via simulation, human demonstrations, and teleoperation) and skill-learning frameworks (imitation and reinforcement learning). Then, based on the overview of the existing data collection paradigm and learning framework, three key challenges re- stricting the development of dexterous robotic manipulation are summarized and discussed.

Index terms

Dexterous Manipulation Multifingered Hands AI-Enabled Robotics

Related papers