UniBYD: A Unified Framework for Learning Robotic Manipulation Across Embodiments Beyond Imitation of Human Demonstrations

TL;DR

UniBYD introduces a unified reinforcement learning framework with a morphological representation to enable diverse robots to learn manipulation tasks beyond imitation, achieving significant success improvements.

Contribution

The paper presents UniBYD, a novel framework combining dynamic reinforcement learning, a unified morphological representation, and a hybrid guidance engine for cross-embodiment robotic manipulation.

Findings

44.08% average success rate improvement over state-of-the-art

Effective transition from imitation to adaptive exploration

Supports diverse robotic hand configurations

Abstract

In embodied intelligence, the embodiment gap between robotic and human hands brings significant challenges for learning from human demonstrations. Although some studies have attempted to bridge this gap using reinforcement learning, they remain confined to merely reproducing human manipulation, resulting in limited task performance. Moreover, current methods struggle to support diverse robotic hand configurations. In this paper, we propose UniBYD, a unified framework that uses a dynamic reinforcement learning algorithm to discover manipulation policies aligned with the robot's physical characteristics. To enable consistent modeling across diverse robotic hand morphologies, UniBYD incorporates a unified morphological representation (UMR). Building on UMR, we design a dynamic PPO with an annealed reward schedule, enabling reinforcement learning to transition from offline-informed imitation of human demonstrations to online-adaptive exploration of policies better adapted to diverse robotic morphologies, thereby going beyond mere imitation of human hands. To address the severe state drift caused by the incapacity of early-stage policies, we design a hybrid Markov-based shadow engine that provides fine-grained guidance to anchor the imitation within the expert's manifold. To evaluate UniBYD, we propose UniManip, the first benchmark for cross-embodiment manipulation spanning diverse robotic morphologies. Experiments demonstrate a 44.08% average improvement in success rate over the current state-of-the-art. Upon acceptance, we will release our code and benchmark.