ManiGaussian: Dynamic Gaussian Splatting for Multi-task Robotic Manipulation
This work addresses the challenge of enabling general intelligent robots to perform multi-task manipulation based on language instructions, representing an incremental advance with a novel method for a known bottleneck in scene dynamics modeling.
The paper tackles the problem of language-conditioned robotic manipulation in unstructured environments by proposing ManiGaussian, a dynamic Gaussian Splatting method that mines scene dynamics via future scene reconstruction, achieving a 13.1% improvement in average success rate over state-of-the-art methods on RLBench tasks.
Performing language-conditioned robotic manipulation tasks in unstructured environments is highly demanded for general intelligent robots. Conventional robotic manipulation methods usually learn semantic representation of the observation for action prediction, which ignores the scene-level spatiotemporal dynamics for human goal completion. In this paper, we propose a dynamic Gaussian Splatting method named ManiGaussian for multi-task robotic manipulation, which mines scene dynamics via future scene reconstruction. Specifically, we first formulate the dynamic Gaussian Splatting framework that infers the semantics propagation in the Gaussian embedding space, where the semantic representation is leveraged to predict the optimal robot action. Then, we build a Gaussian world model to parameterize the distribution in our dynamic Gaussian Splatting framework, which provides informative supervision in the interactive environment via future scene reconstruction. We evaluate our ManiGaussian on 10 RLBench tasks with 166 variations, and the results demonstrate our framework can outperform the state-of-the-art methods by 13.1\% in average success rate. Project page: https://guanxinglu.github.io/ManiGaussian/.