Learning Passive Continuous-Time Dynamics with Multistep Port-Hamiltonian Gaussian Processes
This work addresses the challenge of modeling passive continuous-time dynamics for applications in physics and engineering, representing an incremental improvement with a novel hybrid method.
The paper tackled the problem of learning physically consistent continuous-time dynamics from noisy, irregularly-sampled trajectories by proposing the multistep port-Hamiltonian Gaussian process (MS-PHS GP), which improved vector-field recovery and provided well-calibrated Hamiltonian uncertainty on benchmarks like mass-spring, Van der Pol, and Duffing systems.
We propose the multistep port-Hamiltonian Gaussian process (MS-PHS GP) to learn physically consistent continuous-time dynamics and a posterior over the Hamiltonian from noisy, irregularly-sampled trajectories. By placing a GP prior on the Hamiltonian surface $H$ and encoding variable-step multistep integrator constraints as finite linear functionals, MS-PHS GP enables closed-form conditioning of both the vector field and the Hamiltonian surface without latent states, while enforcing energy balance and passivity by design. We state a finite-sample vector-field bound that separates the estimation and variable-step discretization terms. Lastly, we demonstrate improved vector-field recovery and well-calibrated Hamiltonian uncertainty on mass-spring, Van der Pol, and Duffing benchmarks.