论文arXiv 2307.08336今天

RAYEN框架实现神经网络硬凸约束零违规与千倍加速

RAYEN框架通过避免正交投影与迭代修正，在训练推理全阶段保证硬凸约束满足。施加1K二次约束仅增8ms开销，求解轨迹优化比SOTA快20至7468倍，已在四足机器人 locomotion 策略中验证执行器约束有效性并开源。

AI 整理

本体 SDK 与控制
仿真与训练
评测与安全

本文由 AI 根据公开内容整理，可能有误，请以原文为准：RAYEN: Imposition of Hard Convex Constraints on Neural Networks（arXiv 2307.08336）

核心内容

针对机器人学习中凸约束难以严格满足或计算成本过高的问题，ETH Zurich腿足机器人实验室提出RAYEN框架。该方法支持线性、二次、二阶锥及LMI等多种硬凸约束组合，在不依赖软约束或保守近似的前提下，确保任意输入与权重下约束绝对满足。实测对1K维变量施加1K个二次约束仅增加8ms延迟，300x300 LMI约束在10K维变量上额外开销仅12ms，相比无约束网络几乎无性能损耗。

在工程落地层面，RAYEN用于逼近受约束轨迹优化解时，推理速度较现有最优算法提升20至7468倍且保持近优代价。研究团队将其集成于四足机器人运动控制策略，成功在仿真与真机实验中强制执行执行器物理限制，验证了安全关键场景下的可靠性。相关代码已在 github.com/leggedrobotics/rayen 开源，为具身智能体提供可证明安全的端到端学习范式。

关键看点

硬约束保障：支持线性、SOC及LMI组合约束，训练测试全阶段严格满足，无需正交投影或迭代修正
极低计算开销：1K维变量施加1K二次约束仅增8ms，10K维LMI约束额外耗时12ms，接近无约束网络性能
求解效率跃升：应用于约束轨迹优化问题时，推理速度比SOTA算法快20至7468倍且保持近优代价
真机安全验证：在四足机器人locomotion策略中强制满足执行器约束，仿真与实机实验均证实零违规

工程与产业观察

RAYEN解决了具身智能中安全约束与学习效率的矛盾，其毫秒级硬约束嵌入能力使端到端策略可直接部署于算力受限的机器人本体。该框架将数学规划的安全性融入深度学习，为关节力矩限制、避障等刚性需求提供了可证明合规的工程化方案，显著降低Sim-to-Real迁移风险。

英文摘要 (Abstract)

Despite the numerous applications of convex constraints in Robotics, enforcing them within learning-based frameworks remains an open challenge. Existing techniques either fail to guarantee satisfaction at all times, or incur prohibitive computational costs. This paper presents RAYEN, a framework for imposing hard convex constraints on the output or latent variables of a neural network. RAYEN guarantees constraint satisfaction during both training and testing, for any input and any network weights. Unlike prior approaches, RAYEN avoids computationally expensive orthogonal projections, soft constraints, conservative approximations of the feasible set, and slow iterative corrections. RAYEN supports any combination of linear, convex quadratic, second-order cone (SOC), and linear matrix inequality (LMI) constraints, with negligible overhead compared to unconstrained networks. For instance, it imposes 1K quadratic constraints on a 1K-dimensional variable with only 8 ms of overhead compared to a network that does not enforce these constraints. An LMI constraint with 300x300 dense matrices on a 10K-dimensional variable can be guaranteed with only 12 ms additional overhead. When used in neural networks that approximate the solution of constrained trajectory optimization problems, RAYEN runs 20 to 7468 times faster than state-of-the-art algorithms, while guaranteeing constraint satisfaction at all times and achieving a near-optimal cost. Finally, we demonstrate RAYEN's ability to enforce actuator constraints on a learned locomotion policy by validating constraint satisfaction in both simulation and real-world experiments on a quadruped robot. The code is available at https://github.com/leggedrobotics/rayen

作者：Jesus Tordesillas, Victor Klemm, Jonathan P. How, Marco Hutter