Neural Dynamics Augmented Diffusion Policy

Imitation learning has been proven effective in mimicking demonstrations across various robotic manipulation tasks. However, to develop robust policies, current imitation methods, such as diffusion policy, require training on extensive demonstrations, making data collection labor-intensive. In contrast, model-based planning with dynamics models can effectively cover a sufficient range of configurations using only off-policy data. Yet, without the guidance of expert demonstrations, many tasks are difficult and time-consuming to plan using the dynamics models. Therefore, we take the best of both model learning and imitation learning, and propose neural dynamics augmented imitation learning that covers a large scene configurations with few-shot demonstrations. This method trains a robust diffusion policy in a local support region using few-shot demonstrations and rearranges objects outside this region into it using offline-trained neural dynamics models. Extensive experiments across various tasks in both simulations and real-world scenarios, including granular manipulation, contact-rich task and multi-object interaction task, have demonstrated that trained with only 1 to 30 demonstrations, our proposed method can robustly cover a significantly larger area than the policy trained purely from the demonstrations. Our project page is available at: https://dynamics-dp.github.io.

Authors

• Ruihai Wu
• Haozhe Chen
• Mingtong Zhang 0003
• Haoran Lu
• Yitong Li
• Yunzhu Li

Keywords

• Training
• Imitation learning
• Data collection
• Data models
• Robustness
• Planning
• Robots
• Manipulator dynamics
• Neural Dynamics
• Learning Models
• Dynamic Model
• Neural Model
• Real-world Scenarios
• Manipulation Tasks
• Robot Manipulator
• Robust Policy
• Range Of Configurations
• Expert Demonstrations
• Small Region
• General Solution
• Rigid Body
• Point Cloud
• Multilayer Perceptron
• Random Locations
• Random Orientation
• Objective Conditions
• Graph Neural Networks
• Objects In The Scene
• Number Of Demonstrations
• Policy Learning
• Physics-based Simulation
• Convex Objective
• Perform Complex Tasks
• Coverage Of Space
• Wide Range Of Tasks
• Object Pose