FNODE: Flow-Matching for data-driven simulation of constrained multibody systems
This addresses computational cost and long-term prediction accuracy issues in simulating constrained multibody systems, representing an incremental improvement over prior data-driven approaches.
The paper tackled data-driven modeling of constrained multibody systems by introducing FNODE, a framework that learns acceleration vector fields directly from trajectory data, which outperformed existing methods like MBD-NODE, LSTM, and FCNN across benchmarks such as mass-spring-damper systems and double pendulums, demonstrating good accuracy, generalization, and computational efficiency.
Data-driven modeling of constrained multibody systems faces two persistent challenges: high computational cost and limited long-term prediction accuracy. To address these issues, we introduce the Flow-Matching Neural Ordinary Differential Equation (FNODE), a framework that learns acceleration vector fields directly from trajectory data. By reformulating the training objective to supervise accelerations rather than integrated states, FNODE eliminates the need for backpropagation through an ODE solver, which represents a bottleneck in traditional Neural ODEs. Acceleration targets are computed efficiently using numerical differentiation techniques, including a hybrid Fast Fourier Transform (FFT) and Finite Difference (FD) scheme. We evaluate FNODE on a diverse set of benchmarks, including the single and triple mass-spring-damper systems, double pendulum, slider-crank, and cart-pole. Across all cases, FNODE consistently outperforms existing approaches such as Multi-Body Dynamic Neural ODE (MBD-NODE), Long Short-Term Memory (LSTM) networks, and Fully Connected Neural Networks (FCNN), demonstrating good accuracy, generalization, and computational efficiency.