Search arXiv⌕ Search

arXiv · 2611.09937

Efficient Learning of Differentiable MPC Policies on the GPU via Smoothed ADMM and First-Order Reinforcement Learning

Abstract

Differentiable model predictive control (MPC) enables structured, constraint-aware policies to be optimized end-to-end. However, learning MPC policies faces two bottlenecks: (1) nonsmooth sensitivities at active-set transitions and (2) expensive policy evaluation, since each requires an optimization solve. We first introduce SmoothedTurboMPC, a GPU-parallel differentiable MPC solver combining sequential quadratic programming (SQP) with a log-barrier-smoothed alternating direction method of multipliers (ADMM) inner solver. Its design enables smooth gradients, exact sensitivities at convergence, and efficient execution on GPUs. Second, to efficiently train MPC policies, we pair our solver with first-order reinforcement learning (FORL) using Short-Horizon Actor-Critic (SHAC), which reuses rollout MPC solutions instead of repeatedly re-evaluating them. Comparisons with state of the art solvers show that the approach achieves up to $4.4\times$ higher end-to-end training throughput. On a vehicle racing task, our method achieves higher returns than MPC PPO with a $46\times$ speedup, training on 3.3 million transitions in only 7 minutes. Finally, we use the framework to train a more reliable MPC racing policy via RL with domain randomization, and validate it on a full-scale race car. We integrate our advances into the open-source TurboMPC solver.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Jianghan Zhang, Roland Schwan, Gabriel Bravo-Palacios, John Zhang, Thomas Lew, Brian Plancher. 2026-10-13. Efficient Learning of Differentiable MPC Policies on the GPU via Smoothed ADMM and First-Order Reinforcement Learning. https://arxiv.org/abs/2611.09937

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

Sketch-MoMa: Teleoperation for Mobile Manipulator via Interpretation of Hand-Drawn Sketches

To use assistive robots in everyday life, a remote control system with common devices, such as 2D devices, is helpful to control the robots anytime and anywhere as intended. Hand-drawn sketches are one of the intuitive ways to control robots with 2D devices. However, since similar sketches have different intentions from scene to scene, existing work needs additional modalities to set the sketches' semantics. This requires complex operations for users and leads to decreasing usability. In this paper, we propose Sketch-MoMa, a teleoperation system using the user-given hand-drawn sketches as instructions to control a robot. We use Vision-Language Models (VLMs) to understand the user-given sketches superimposed on an observation image and infer drawn shapes and low-level tasks of the robot. We utilize the sketches and the generated shapes for recognition and motion planning of the generated low-level tasks for precise and intuitive operations. We validate our approach using state-of-the-art VLMs with 7 tasks and 5 sketch shapes. We also demonstrate that our approach effectively specifies the detailed motions, such as how to grasp and how much to rotate. Moreover, we show the competitive usability of our approach compared with the existing 2D interface through a user experiment with 14 participants.

cs.RO↗

Evolving Drone Morphologies for Agile Flight

The design of multicopter drones has remained almost the same since their inception. While conventional designs work well in many cases, they may not be optimal in specific tasks. This paper revisits rotary drone design by exploring which body morphologies are best suited for different objectives and constraints. An evolutionary algorithm is used to produce drone morphologies that performs better than standard configurations at thrust-to-weight ratio, maneuverability and size. To generate a range of drones with performance trade-offs between them, the non-dominated sorting genetic algorithm II is used. The evolved drone bodies outperform a standard multicopters in at least one of the three objectives. The Pareto front extrema show improvement of 489.2%, 23% and 4.8% in maneuverability, thrust-to-weight ratio and size, respectively, when compared to a standard quadcopter. The improvement in maneuverability is attributed to the tilt angles of the motors, while the increase in thrust-to-weight ratio is due to the higher number of propellers. Experiments show that the performance improvement transfers to a real world drone racing task, where the evolved drone outperforms the standard drone in terms of number of gates passed,flight speed and lap times.

cs.RO↗

ManiTaskGen: A Comprehensive Task Generator for Benchmarking and Improving Vision-Language Agents on Embodied Decision-Making

Building embodied agents capable of accomplishing arbitrary tasks is a core objective towards achieving embodied artificial general intelligence (E-AGI). While recent work has advanced such general robot policies, their training and evaluation are often limited to tasks within specific scenes, involving restricted instructions and scenarios. Existing benchmarks also typically rely on manual annotation of limited tasks in a few scenes. We argue that exploring the full spectrum of feasible tasks within any given scene is crucial, as they provide both extensive benchmarks for evaluation and valuable resources for agent improvement. Towards this end, we introduce ManiTaskGen, a novel system that automatically generates comprehensive, diverse, feasible mobile manipulation tasks for any given scene. The generated tasks encompass both process-based, specific instructions (e.g., "move object from X to Y") and outcome-based, abstract instructions (e.g., "clear the table"). We apply ManiTaskGen to both simulated and real-world scenes, demonstrating the validity and diversity of the generated tasks. We then leverage these tasks to automatically construct benchmarks, evaluating high-level embodied decision-making through a grounded skill interface for agents built upon existing vision-language models (VLMs). Furthermore, we propose a simple yet effective method that utilizes ManiTaskGen tasks to enhance embodied decision-making. Overall, this work presents a cross-scene task generation framework facilitating both benchmarking and improvement of embodied decision-making agents. Our code is available at https://manitaskgen.github.io.

cs.RO↗