[TOC]

本文为《无人驾驶车辆局部路径规划的时间一致性与鲁棒性研究》的阅读笔记.

学校: 北理

作者: 姜岩, 王琦等

关键词: 局部规划, 时间一致性(temporary consistency), 鲁棒性(robustness)

[TOC]

本文为论文《Safe Trajectory Generation for Complex Urban Environments Using Spatio-temporal Semantic Corridor》的阅读后总结.
作者:Wenchao Ding,Lu Zhang 等
学校:香港科技大学
主题: frenet 坐标系下建图后, 利用 QP 进行局部轨迹规划.
论文网址: https://ieeexplore.ieee.org/document/8740885

[TOC]

本文为论文《EPSILON: An Efficient Planning System for Automated Vehicles in Highly Interactive Environments》的阅读后总结.
作者:Wenchao Ding,Lu Zhang 等
学校:香港科技大学
主题: 将预测和决策融合, 使用 POMDP 的概念包装决策树, 实现城市拥堵路段的自主换道. 局部轨迹规划可以参考他们之前的论文,解读博文链接.
论文网址: https://arxiv.org/pdf/2108.07993.pdf
开源地址: https://github.com/HKUST-Aerial-Robotics/EPSILON

[TOC]
对于目前的基于状态机或者部分融入优化cost function里的决策算法,我觉得会很容易走到瓶颈.于是查一下学术前沿的一些决策算法.总结于此文.论文地址

[TOC]
本文记录一下《TDR-OBCA: A Reliable Planner for Autonomous Driving in Free-Space Environment》论文的阅读与公式推导.
所的出来的模型为 Apollo open space planner 代码依据.

论文要点:

改进 H-OBCA(参见研读推导论文博客)模型:

1. 2 个 warm start: Temporal and Dual variable warm starts.
2. 改进模型, Reformulates the problem

[TOC]
本文记录一下《Optimization-Based Collision Avoidance》的阅读与公式推导.
此论文是 Apollo open space planner 中的模型原型.

论文要点:

1. 将控制体与障碍物描述成凸集(convex sets such as polytopes or ellipsoids (or can be decomposed into a finite union of such convex sets)), 然后通过拉格朗日对偶变换成相应的凸优化形式.
2. 即便在计算失败碰撞时也要尽可能地侵入过少, 引入 signed distance 概念.
3. 仿真验证在无人机飞行避障/汽车倒车环境有效(使用 A* 搜索作为 warm start).

Abstract

三大特点:
(1) The top layer of the system is a multilane strategy that handles lane-change scenarios by comparing lane-level trajectories computed in parallel.
(2) Inside the lane-level trajectory generator, it iteratively solves path and speed optimization based on a Frenet frame.
(3) For path and speed optimization, a combination of dynamic programming and spline-based quadratic programming is proposed to construct a scalable and easy-to-tune framework to handle traffic rules, obstacle decisions and smoothness simultaneously.

Basic info

Author: Yu Zhang
Date: 2018
Juounal: IEEE Acess
Institute: 北理

Abstract

共四方面贡献:

1. trajectory planning framework: 克服 geometry constraints(free space), nonholonomic constraints(vehicle kinematics and dynamics) and dynamics constraints, task constraints (requirements to visit certain intermediate targets)实现 curvature continuous , kinodynamically feasible, smooth and collision-free trajectories in real-time.
2. derivative-free global path modification algorithm to extract high-order state information in free space for state sampling.
3. multi-phase deterministic state space sampling method that is able to approximate complex maneuvers.
4. car footprint approximation strategy and a two-phase collision checking routine.
   后面这些名词概念都会得到详细解释.