Embodiment Randomization for Cross Embodiment Navigation

We present Embodiment Randomization, a simple, inexpensive, and intuitive technique for training robust behavior policies that can be transferred to multiple robot embodiments. While prior works require real-world data from multiple robots, or complex algorithmic adjustments to address the challenge of embodiment generalization, our approach leverages the power of simulation and large-scale reinforcement learning and can be easily integrated within existing policy learning methods. We show that policies trained with embodiment randomization implicitly perform system identification, enabling them to adapt to new embodiments during deployment. Our approach not only shows significant improvements in adapting to novel robot configurations, but also in generalizing from simulation to reality and contending with real-world perturbations, highlighting the potential of embodiment randomization in creating versatile and adaptable robotic navigation policies.

Authors

• Pranav Putta
• Gunjan Aggarwal
• Roozbeh Mottaghi
• Dhruv Batra
• Naoki Yokoyama
• Joanne Truong
• Arjun Majumdar

Keywords

• Training
• Learning systems
• Visualization
• Navigation
• Perturbation methods
• Reinforcement learning
• System identification
• Reliability
• Probes
• Robots
• Real-world Data
• Policy Learning
• Training Policy
• Multiple Robots
• Behavior Policy
• Inexpensive Technique
• Robot Configuration
• Step Size
• Path Length
• Data Augmentation
• Simulation Performance
• Domain Adaptation
• Real-world Environments
• Large Parameter
• Machine Vision
• Graph Topology
• Discrete Action
• Navigation Task
• Beginning Of Episode
• Policy Transfer
• Single Policy
• Start Location
• Recovery Error
• Navigation Performance
• Largest Drop
• Intermediate Representation
• Simulated Robot