运动规划与控制
运动规划与控制
Section titled “运动规划与控制”运动规划、轨迹优化与底盘/机械臂控制执行。
类型 Topics · 更新 2026-07-19
本 Topic 要解决什么
Section titled “本 Topic 要解决什么”本Topic聚焦自动驾驶中运动规划与控制核心问题：如何在动态不确定环境中生成安全可行的轨迹，并实现高精度实时跟踪。涵盖经典采样/优化规划、模型预测控制（MPC）路径跟踪，以及学习式规划（如模仿学习）的融合。目标是连接高层决策与底层执行，处理车辆动力学约束、多障碍避障、实时性与闭环性能，支撑城市驾驶场景（参考nuPlan等）。解决规则方法可解释但泛化弱、纯学习易分布偏移的矛盾，推动classical+learning混合范式落地。
非目标 / 边界
Section titled “非目标 / 边界”- 纯感知建图或端到端无显式规划模块
- 仅海洋/无人机非道路车辆专项
- 底层电机驱动硬件细节或纯仿真无闭环评估
- 高层任务规划或长期行为预测独立专题
foundations— 需掌握车辆运动学/动力学基本模型与优化控制基础，以便理解MPC约束与轨迹可行性motion-prediction— 规划依赖周边智能体轨迹预测输入，需理解交互不确定性如何传入规划ad-prediction-planning-control— 本Topic细化规划-控制链路，承接整体预测-规划-控制栈架构
| 术语 | 含义 | 常见混淆 |
|---|---|---|
| 模型预测控制 (MPC) | 在有限时域内滚动优化控制序列，利用系统模型预测未来状态并处理约束，实现轨迹跟踪 | 易与LQR混淆：MPC显式处理硬约束与多目标，滚动优化而非一次求解 |
| 轨迹跟踪 (Trajectory Tracking) | 使车辆实际状态（位置、速度、航向）精确跟随给定参考轨迹的控制问题 | 与路径跟踪区别：轨迹含时间参数，路径仅几何；跟踪强调动态响应 |
| 采样式运动规划 | 通过随机或确定性采样探索配置空间，构建路径树/图（如RRT系列）寻找可行/最优路径 | 易与网格搜索混淆：采样适合高维连续空间，但有概率完备性而非确定性 |
| 路径规划 vs 轨迹规划 | 路径规划输出几何曲线；轨迹规划进一步分配时间/速度形成时空曲线 | 常混用：完整规划需从path到trajectory再tracking |
| 模仿学习规划 | 从专家轨迹数据学习规划策略，直接输出未来轨迹（如PLUTO等） | 易与强化学习混淆：模仿偏监督学习，易shortcut与分布偏移；RL需奖励设计 |
| 直接横摆力矩控制 (DYC) | 通过差动制动/驱动产生横摆力矩辅助转向，提升轨迹跟踪与稳定性 | 与主动前轮转向区别：DYC作用于纵向力分配，适合极限工况 |
方法谱系与时间线
Section titled “方法谱系与时间线”早期以采样（RRT）与优化（MPC）为主，强调可解释与约束满足；中期分层采样+优化提升效率与质量；近年学习式（模仿/混合）崛起，在nuPlan等闭环基准上突破规则方法，但引入安全与可验证挑战。当前趋势为数据-物理模型混合与实时近似MPC，兼顾性能与保障。
- (2009) watch — Real-Time Motion Planning With Applications to Autonomous Urban Driving
- (2010) watch — Incremental Sampling-based Algorithms for Optimal Motion Planning
- (2014) watch — Sampling-Based Robot Motion Planning: A Review
- (2015) recent — A Review of Motion Planning Techniques for Automated Vehicles
- (2015) needs-review — Trajectory-Tracking Control of Mobile Robot Systems Incorporating Neural-Dynamic Optimized Model Predictive Approach
- (2016) foundational — Path Planning and Tracking for Vehicle Collision Avoidance Based on Model Predictive Control With Multiconstraints
- (2018) watch — Hierarchical Trajectory Planning of an Autonomous Car Based on the Integration of a Sampling and an Optimization Method
- (2019) watch — A Motion Planning and Tracking Framework for Autonomous Vehicles Based on Artificial Potential Field Elaborated Resistance Network Approach
- (2021) foundational — Review on model predictive control: an engineering perspective
- (2021) needs-review — A Survey of Deep Reinforcement Learning Algorithms for Motion Planning and Control of Autonomous Vehicles
- (2022) foundational — Model predictive path tracking control for automated road vehicles: A review
- (2022) recent — Mpc-based motion planning and control enables smarter and safer autonomous marine vehicles: perspectives and a tutorial survey
- (2024) recent — A direct yaw moment control frame through model predictive control considering vehicle trajectory tracking performance and handling stability for autonomous driving
- (2024) recent — Path planning algorithms in the autonomous driving system: A comprehensive review
- (2024) recent — PLUTO: Pushing the Limit of Imitation Learning-based Planning for Autonomous Driving
- (2025) watch — Trajectory Tracking Control for Autonomous Driving Adaptive Cruise Control by Fusion of Model Predictive Control and Deep Reinforcement Learning
- (2025) watch — Hybrid Data- and Physical Model-driven Safe and Intelligent Motion Planning and Control for Autonomous Vehicles
- (2026) watch — Real-time Gaussian Process based Approximate Model Predictive Trajectory Tracking Control for Autonomous Vehicles
- Model predictive path tracking control for automated road vehicles: A review（精读）— 系统综述自动驾驶道路车辆MPC路径跟踪方法，奠定经典控制基础与工程视角
- Review on model predictive control: an engineering perspective（精读）— 从工程角度全面回顾MPC原理、实现与挑战，支撑轨迹跟踪控制理解
- PLUTO: Pushing the Limit of Imitation Learning-based Planning for Autonomous Driving（中文笔记）— 学习式规划代表作，通过架构、辅助损失与对比模仿学习突破局限，在nuPlan闭环超越规则规划器
推荐阅读路径
Section titled “推荐阅读路径”- 先读两篇MPC综述建立经典控制与路径跟踪基础
- 再读运动规划技术综述与采样式方法（如RRT相关），理解规划全景
- 最后精读PLUTO与近期混合/MPC融合论文，对比学习与经典闭环表现，结合nuPlan思考
| Paper | 输入 | 输出 | 表示 | 训练目标 |
|---|---|---|---|---|
| PLUTO: Pushing the Limit of Imitation Learning-based Planning for Autonomous Driving | 场景观察/地图/历史轨迹 | 未来轨迹规划 | 横向-纵向感知查询融合架构 | 对比模仿学习+辅助损失，闭环超越规则规划器 |
| Model predictive path tracking control for automated road vehicles: A review | 参考路径/轨迹与车辆状态 | 控制输入序列（转向/加速度） | 车辆动力学模型+约束优化 | 最小化跟踪误差并满足约束，滚动优化 |
| Path Planning and Tracking for Vehicle Collision Avoidance Based on Model Predictive Control With Multiconstraints | 障碍信息与参考 | 避碰路径/控制 | 多约束MPC | 碰撞避免同时轨迹跟踪 |
| Sampling-Based Robot Motion Planning: A Review | 配置空间/障碍 | 可行路径 | 采样树/图 | 概率完备路径搜索 |
工程实现与复现风险
Section titled “工程实现与复现风险”- 实时性：MPC优化与高维采样计算成本高，需近似或分层
- 模型失配：动力学简化导致跟踪误差或失稳
- 安全验证：学习规划分布偏移与黑盒难保证闭环安全
- 约束处理：多约束下可行性与舒适性权衡
- sim-to-real与数据偏差：nuPlan等基准到真实车辆迁移
- 如何系统融合物理模型MPC与数据驱动规划，同时保持可解释与实时
- 学习式规划在复杂交互与长尾场景的鲁棒闭环保障机制
- 近似MPC（如高斯过程）的理论误差界与实际车辆验证
- 多约束（舒适、能耗、法规）下的统一优化框架
Topic 自测清单
Section titled “Topic 自测清单”- MPC路径跟踪与传统PID跟踪的核心区别是什么？如何处理硬约束？
- 采样式规划（如RRT*）的概率完备性含义及与最优性的关系？
- PLUTO如何通过架构与训练框架缓解模仿学习的shortcut与分布偏移问题？
- 轨迹规划与路径规划的区别，以及完整系统中如何衔接跟踪控制？
- 在碰撞避免场景中，多约束MPC通常如何建模障碍与车辆动力学？
选篇附录（registry）
Section titled “选篇附录（registry）”本 Topic registry 入选 18 篇（foundational/recent/watch 见 papers.yml）。
Planning and Control for Autonomous Vehicles](/content/literature-doi-10-4271-epr2025014/) | 2025 | auto score=40 || recent | PLUTO: Pushing the Limit of Imitation Learning-based Planning for Autonomous Driving | 2024 | PLUTO is a leading learned motion planner for urban AD |
论文池（FRW）
Section titled “论文池（FRW）”来自 research/papers.yml 的正式选题配额：基础必读 3 · 近期重要 5 · 观察 5。
可学习条目需本地 PDF 或已有精读/中文笔记；无正文资产的选题标为 待获取 PDF，不作为学习路径必读链接。
基础必读（选题 3 · 可学习 0 · 待获取 PDF 3）
Section titled “基础必读（选题 3 · 可学习 0 · 待获取 PDF 3）”本档尚无可学习正文（PDF/中文笔记）；下列为待获取选题。
- Review on model predictive control: an engineering perspective (2021) · 待获取 PDF
- Model predictive path tracking control for automated road vehicles: A review (2022) · 待获取 PDF
- Path Planning and Tracking for Vehicle Collision Avoidance Based on Model Predictive Control With Multiconstraints (2016) · 待获取 PDF
近期重要（选题 5 · 可学习 1 · 待获取 PDF 4）
Section titled “近期重要（选题 5 · 可学习 1 · 待获取 PDF 4）”- PLUTO: Pushing the Limit of Imitation Learning-based Planning for Autonomous Driving (2024) · 固定 · 中文笔记
- A Review of Motion Planning Techniques for Automated Vehicles (2015) · 待获取 PDF
- A direct yaw moment control frame through model predictive control considering vehicle trajectory tracking performance and handling stability for autonomous driving (2024) · 待获取 PDF
- Mpc-based motion planning and control enables smarter and safer autonomous marine vehicles: perspectives and a tutorial survey (2022) · 待获取 PDF
- …另有 1 篇待获取 PDF
观察清单（选题 8 · 可学习 3 · 待获取 PDF 5）
Section titled “观察清单（选题 8 · 可学习 3 · 待获取 PDF 5）”- Real-time Gaussian Process based Approximate Model Predictive Trajectory Tracking Control for Autonomous Vehicles (2026) · 中文笔记
- Hybrid Data- and Physical Model-driven Safe and Intelligent Motion Planning and Control for Autonomous Vehicles (2025) · 中文笔记
- Trajectory Tracking Control for Autonomous Driving Adaptive Cruise Control by Fusion of Model Predictive Control and Deep Reinforcement Learning (2025) · 中文笔记
- A Motion Planning and Tracking Framework for Autonomous Vehicles Based on Artificial Potential Field Elaborated Resistance Network Approach (2019) · 待获取 PDF
- Sampling-Based Robot Motion Planning: A Review (2014) · 待获取 PDF
- Real-Time Motion Planning With Applications to Autonomous Urban Driving (2009) · 待获取 PDF
- …另有 2 篇待获取 PDF
成熟度 seed · 内容数 19 · 论文池 16