Accelerating Sampling-Based Control via Learned Linear Koopman Dynamics
For roboticists needing real-time control of systems with complex nonlinear dynamics, this work offers a computationally efficient MPPI variant that maintains performance without requiring an analytical model.
This paper accelerates model predictive path integral (MPPI) control by replacing nonlinear dynamics with a learned linear deep Koopman operator model, achieving similar control performance to MPPI with true dynamics while substantially reducing computational cost for real-time robotic control.
This paper presents an efficient model predictive path integral (MPPI) control framework for systems with complex nonlinear dynamics. To improve the computational efficiency of classic MPPI while preserving control performance, we replace the nonlinear dynamics used for trajectory propagation with a learned linear deep Koopman operator (DKO) model, enabling faster rollout and more efficient trajectory sampling. The DKO dynamics are learned directly from interaction data, eliminating the need for analytical system models. The resulting controller, termed MPPI-DK, is evaluated in simulation on pendulum balancing and surface vehicle navigation tasks, and validated on hardware through reference-tracking experiments on a quadruped robot. Experimental results demonstrate that MPPI-DK achieves control performance close to MPPI with true dynamics while substantially reducing computational cost, enabling efficient real-time control on robotic platforms.