|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/autonomous-robots-nonholonomic-motion-planning-algorithms/
课程评论：没有评论
**课程名称：** 自动驾驶机器人：非完整运动规划算法 (Autonomous Robots: Nonholonomic Motion Planning Algorithms) **课程概述：** 本课程是“抽样基础运动规划算法导论”的上、下部课程。本课程将涵盖大量新内容，无需提前完成前置课程。课程主要内容包括： * **学习仿射变换和向量旋转矩阵。** * **推导Dubins路径，生成平滑的轨迹。** * **将Dubins路径与抽样基础运动规划算法（RRT和RRT*）相结合。** * **完成4个互动作业。** **详细介绍：** 运动规划是工程学的一个分支，旨在计算机器人从起点到终点的无障碍路径，例如谷歌地图或苹果地图的应用。当今世界上的大多数车辆（如汽车、船只、飞机）都属于非完整性车辆，这意味着它们相对于可用空间具有有限的自由度。以汽车在相邻停车位之间移动为例，汽车不能“侧滑”到第二个停车位，它必须倒车并转入，或者绕一个圆形路径进入。 本课程将引导你，给定地图上的两个任意位置和朝向，找到满足车辆速度约束的平滑轨迹。这可以通过Dubins路径实现。你将学习如何从第一原理推导Dubins路径，并通过作业进行实现和测试。之后，你将 lernen 如何将此理论与抽样基础运动规划算法相结合。接下来的两个作业将涉及使用RRT和RRT\*结合Dubins路径寻找路径。最后一个作业将针对一个机器人信息有限的真实地图场景，使用增量RRT结合Dubins路径进行规划。 **学习要求：** * **编程语言：** Python（推荐），Numpy和Matplotlib。 * **基础知识：** 对标量和向量有基本理解。 * **其他语言：** 可以使用其他编程语言完成课程，但需要自行将作业转换为相应语言。
Important Note:This course is a Part 2 to my free course Introduction to Sampling Based Motion Planning Algorithms.However, you do not need to finish the free course first, since I cover all of the free course material (plus a whole lot more!) within this course. With that said, if you already have completed Introduction to Sampling Based Motion Planning Algorithms, then you will have an advantage.HighlightsLearn affine transformations and vector rotation matrices.Derivation of the Dubins Path to generate smooth trajectories given two points on a map.Combine Dubins Path with motion planning RRT and RRT* algorithms.4 interactive assignments.Description Motion planning is a field of engineering dealing with calculating a path from a start to a goal location whilst avoiding obstacles, for example using Google or Apple Maps. The vast majority of vehicles in today's world i.e (cars, boats, planes) are non holonomic, meaning they have limited degrees of freedom of movement with respect to the available space. For instance, think of shifting a car between adjacent parking spots. The car cannot 'slide' sideways into the second parking spot. It has to reverse and turn in to the spot, or go around in a circular path. Given two arbitrary positions and headings on a map, the task is to find a smooth trajectory which satisfies the speed constraints of the vehicle. This is accomplished by using the Dubins Path. In this course, you will learn how to derive the Dubins Path from first principles and also implement and test the method via an assignment. Next, you will learn how to combine this theory with sampling based motion planning algorithms. The next two assignments will involve finding a path using RRT and RRT* with Dubins Path. Lastly, the final assignment will involve incremental RRT with Dubins Path for a realistic roadmap scenario, where the vehicle has limited information.RequirementsPython, Numpy and MatplotlibBasic understanding of scalars and vectorsYou could alternatively complete the course using another programming language, however in this case you will have to transcribe the assignments on your own, in that specific language.
