Strengthening Generative Robot Policies through Predictive World Modeling
This addresses the challenge of robust robotic control for manipulation tasks, representing an incremental improvement over existing imitation learning methods.
The paper tackles the problem of improving robotic manipulation policies by introducing generative predictive control (GPC), which combines a diffusion-based policy with a predictive world model and online planning, resulting in consistent outperformance over behavior cloning across various tasks in simulation and real-world settings.
We present generative predictive control (GPC), a learning control framework that (i) clones a generative diffusion-based policy from expert demonstrations, (ii) trains a predictive action-conditioned world model from both expert demonstrations and random explorations, and (iii) synthesizes an online planner that ranks and optimizes the action proposals from (i) by looking ahead into the future using the world model from (ii). Across a variety of robotic manipulation tasks, we demonstrate that GPC consistently outperforms behavior cloning in both state-based and vision-based settings, in simulation and in the real world.