Scalable Differentiable Physics for Learning and Control
This addresses the problem of inefficient differentiable physics for researchers and practitioners in robotics and simulation, offering a scalable solution that is incremental but with significant performance gains.
The paper tackles the scalability limitations of differentiable physics solvers by developing a framework that supports many objects and interactions using meshes and sparse contact handling, reducing memory and computation by up to two orders of magnitude compared to particle-based methods and outperforming baselines by at least an order of magnitude in inverse problems and control.
Differentiable physics is a powerful approach to learning and control problems that involve physical objects and environments. While notable progress has been made, the capabilities of differentiable physics solvers remain limited. We develop a scalable framework for differentiable physics that can support a large number of objects and their interactions. To accommodate objects with arbitrary geometry and topology, we adopt meshes as our representation and leverage the sparsity of contacts for scalable differentiable collision handling. Collisions are resolved in localized regions to minimize the number of optimization variables even when the number of simulated objects is high. We further accelerate implicit differentiation of optimization with nonlinear constraints. Experiments demonstrate that the presented framework requires up to two orders of magnitude less memory and computation in comparison to recent particle-based methods. We further validate the approach on inverse problems and control scenarios, where it outperforms derivative-free and model-free baselines by at least an order of magnitude.