A First Principles Approach for Data-Efficient System Identification of Spring-Rod Systems via Differentiable Physics Engines
This addresses data-efficient and explainable system identification for physics-based simulations, particularly in robotics or engineering domains, though it appears incremental as it builds on differentiable physics engines with specific optimizations.
The paper tackles system identification for complex spring-rod assemblies by proposing a differentiable physics engine that modularizes governing equations and reduces dimensions, enabling efficient parameter learning with linear regression and explainable results. It demonstrates reduced training data needs and avoids iterative processes, showing efficacy on tensegrity systems like NASA's icosahedron.
We propose a novel differentiable physics engine for system identification of complex spring-rod assemblies. Unlike black-box data-driven methods for learning the evolution of a dynamical system and its parameters, we modularize the design of our engine using a discrete form of the governing equations of motion, similar to a traditional physics engine. We further reduce the dimension from 3D to 1D for each module, which allows efficient learning of system parameters using linear regression. As a side benefit, the regression parameters correspond to physical quantities, such as spring stiffness or the mass of the rod, making the pipeline explainable. The approach significantly reduces the amount of training data required, and also avoids iterative identification of data sampling and model training. We compare the performance of the proposed engine with previous solutions, and demonstrate its efficacy on tensegrity systems, such as NASA's icosahedron.