aiqbench analysis
This paper introduces ZipDrift, a generative model that resolves multimodal optimization challenges in short-horizon Model Predictive Control (MPC) for robotic overtaking tasks. By training a drifting generator on cost corrections rather than differentiating through the optimizer, it avoids the instability and inconsistency inherent in deterministic regression approaches when expert labels are ambiguous. The method achieves full collision-free performance in kinematic benchmarks where standard MPC fails significantly, demonstrating a practical advance in learning-based control systems.
ZipDrift: Generating Multimodal Short-Horizon MPC Costs with One-Step Drifting Models
arXiv:2611.05793v1 Announce Type: new Abstract: Short-horizon model predictive control (MPC) is fast but myopic. ZipMPC compensates by learning a context-dependent correction of the short-horizon cost that imitates a long-horizon expert. We observe that the optimal correction is often multimodal: when a slower opponent can be passed on either side, good corrections of opposite sign coexist, and the expert's choice between them hinges on tiny cost differences. Deterministic regression then averages modes or inherits inconsistent expert tie-breaking. We propose ZipDrift, which trains a one-step drifting generator over MPC cost corrections, tilted toward low closed-loop cost and learned only from cost evaluations of candidate corrections, without differentiating through the optimiser. A few generated candidates are ranked by a learned critic, and the model-based MPC then enforces dynamics and constraints. In a kinematic overtaking benchmark derived from ZipMPC, 53% of contexts are bimodal and the manual short-horizon MPC crashes in 28.5% of test episodes. With an eight-dimensional stagewise correction, ZipDrift is collision-free in all 600 test episodes and matches the best evaluated candidate, whereas an idealised ZipMPC-style regressor crashes in 4.0% (16.0% with inconsistent expert labels) and critic-ranked random candidates with the same budget in 2.0%.
Abstract:Short-horizon model predictive control (MPC) is fast but myopic. ZipMPC compensates by learning a context-dependent correction of the short-horizon cost that imitates a long-horizon expert. We observe that the optimal correction is often multimodal: when a slower opponent can be passed on either side, good corrections of opposite sign coexist, and the expert's choice between them hinges on tiny cost differences. Deterministic regression then averages modes or inherits inconsistent expert tie-breaking. We propose ZipDrift, which trains a one-step drifting generator over MPC cost corrections, tilted toward low closed-loop cost and learned only from cost evaluations of candidate corrections, without differentiating through the optimiser. A few generated candidates are ranked by a learned critic, and the model-based MPC then enforces dynamics and constraints. In a kinematic overtaking benchmark derived from ZipMPC, 53% of contexts are bimodal and the manual short-horizon MPC crashes in 28.5% of test episodes. With an eight-dimensional stagewise correction, ZipDrift is collision-free in all 600 test episodes and matches the best evaluated candidate, whereas an idealised ZipMPC-style regressor crashes in 4.0% (16.0% with inconsistent expert labels) and critic-ranked random candidates with the same budget in 2.0%.
6 pages, 2 figures, 4 tables. Preliminary simulation study
Robotics (cs.RO); Machine Learning (cs.LG); Systems and Control (eess.SY)
arXiv:2611.05793 [cs.RO]
(or arXiv:2611.05793v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2611.05793 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 297 words