# SmoothedTurboMPC实现GPU并行可微MPC，赛车策略训练提速46倍

> 原文链接：https://robotics.hellobike.com/papers/2026-11-10-2611-09937-16dda81f
> 频道：papers | 更新日期：2026-11-10 | 标签：auto, arxiv
> 核心摘要：提出SmoothedTurboMPC求解器结合SHAC算法，解决可微MPC梯度非光滑与评估昂贵瓶颈。端到端训练吞吐量提升4.4倍，赛车任务训练330万步仅耗时7分钟，较MPC PPO提速46倍，已集成至开源TurboMPC并在全尺寸赛车上验证。

# SmoothedTurboMPC实现GPU并行可微MPC，赛车策略训练提速46倍

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[Efficient Learning of Differentiable MPC Policies on the GPU via Smoothed ADMM and First-Order Reinforcement Learning](https://arxiv.org/abs/2611.09937)（arXiv 2611.09937）

## 核心内容

针对可微模型预测控制（MPC）在主动集切换时梯度非光滑及策略评估计算昂贵的两大瓶颈，研究团队提出SmoothedTurboMPC框架。该方法将序列二次规划（SQP）与对数障碍平滑ADMM内层求解器结合，支持GPU并行计算，确保收敛时梯度精确且光滑。配合短视域Actor-Critic（SHAC）算法复用rollout解，避免重复优化求解，显著降低训练开销。

实验表明，该方案端到端训练吞吐量较现有SOTA求解器提升4.4倍。在车辆 racing 任务中，仅用7分钟完成330万步过渡训练，回报优于MPC PPO且速度提升46倍。通过引入域随机化增强策略鲁棒性，并成功部署于全尺寸赛车进行实车验证。相关技术已集成至开源项目 `TurboMPC`，为具身智能体实时控制提供高效可微优化基础设施。

## 关键看点

- **平滑ADMM求解**：结合SQP与log-barrier平滑ADMM，消除主动集切换梯度突变，支持GPU并行加速
- **训练效率突破**：采用SHAC算法复用MPC解，端到端吞吐量提升4.4倍，赛车任务训练提速46倍
- **实车验证落地**：基于域随机化训练高可靠MPC策略，并在真机全尺寸赛车上完成物理验证
- **开源资产集成**：核心算法已封装至开源求解器 `TurboMPC`，支持可微MPC策略端到端学习

## 工程与产业观察

该工作解决了可微MPC在具身控制中梯度不稳定与计算慢的工程痛点，通过算法-硬件协同设计将训练效率提升两个数量级。开源 `TurboMPC` 为机器人开发者提供了兼顾约束处理与端到端学习的高效工具，对高性能移动机器人及自动驾驶的实时策略优化具有重要参考价值。

## 英文摘要 (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.

作者：Jianghan Zhang, Roland Schwan, Gabriel Bravo-Palacios, John Zhang, Thomas Lew, Brian Plancher
:::

## 原始参考来源 (Sources)

- [arXiv 2611.09937：Efficient Learning of Differentiable MPC Policies on the GPU via Smoothed ADMM and First-Order Reinforcement Learning](https://arxiv.org/abs/2611.09937)