什么是好的机器人模拟器?

发布于 2024-08-16 05:33:10 字数 341 浏览 2 评论 0原文

所以我目前正在做一些机器人研究,我们在实验室里有一个定制的手臂,我们想模拟它。我做了一些谷歌搜索和维基百科搜索,似乎有相当多的开源和专有应用程序可以满足我的需要,例如 Gazebo、LpzRobots、Webots、RoboLogix 等。我们的主要研究工具是 C++ 和网络接口。

我的问题是,有没有人有处理机器人模拟器的经验?你能建议我应该跳入其中​​哪一个,或者我是否应该拿起并开始编写自己的模拟器划痕?我对它的主要用途是运动规划测试(因此我希望能够精确控制机器人的位置,而不必担心动力学)和自定义运动链<的控制器动力学测试(正好相反) /strong> 机器人。

So I'm currently doing some robotics research, and we have a custom built arm in the lab that we would like to simulate. I've done a little googling and wikipedia searching, and there seems to be a fair amount of open source and proprietary applications for doing just what I need, such as Gazebo, LpzRobots, Webots, RoboLogix, and others. Our primary research tools are in C++ and interface over network.

My question is, does anyone out there have any experience in dealing with robotics simulators? Can you recommend which if any of these I should jump for, or whether I should just pick up and start writing my own from scratch? My primary use for it will be motion planning testing (so I'd like to be able to control the position of the robot exactly without worrying about dynamics) and controller dynamics testing (just the opposite) for a custom kinematic chain robot.

如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

扫码二维码加入Web技术交流群

发布评论

需要 登录 才能够评论, 你可以免费 注册 一个本站的账号。

评论(4

心在旅行 2024-08-23 05:33:10

Microsoft Robotics 拥有 MS Robotics Studio,这是一整套工具适用于任何类型的机器人...无人机、UGV、机械手等。

此类工具之一是基于 AGEIA 物理引擎的 VSE(视觉模拟环境),它不仅可以模拟机器人的运动学,还可以模拟传感器和环境。

他们甚至可能有您正在使用的手臂的 3D 模型,例如,我知道他们有 KUKA LBR3 的模型。

替代文本
(来源:microsoft.com)

Microsoft Robotics have the MS Robotics Studio which is a full set of tools for any type of robot... UAVs, UGVs, Manipulators, etc.

One of such tools is the VSE (Visual Simulation Environment) based on the AGEIA Physics engine that allows simulation of not only the kinematics of your robot, but also simulation of sensors and of the environment.

They might even have the 3D model of the arm you are using, I know that they have one for the KUKA LBR3 for example.

alt text
(source: microsoft.com)

云雾 2024-08-23 05:33:10

除非迫不得已,否则我绝不会建议您编写自己的模拟器。这不是一个很好的学习经验,而且永远不会像那些经过数千人多年来改进的经验那样有效。

话虽如此,我获得的积极经验是 Gazebo 和开放动态引擎(如果您使用 Matlab 进行模拟)。我绝对推荐将 Gazebo 与 ROS 结合使用。它们都变得更加标准并不断改进。您可以根据预先存在的模拟进行模拟,例如不稳定的机器人(用于两轮差速驱动平台)和 PR2(用于人形平台)。如果您只想进行碰撞检查和运动学,那么我建议使用 rvizurdf 格式。非常容易上手并且经过充分测试。

顺便说一下,我过去使用过Microsoft Robotics Developers Studio,它不如ROS。我知道他们有一个新版本可以解决很多以前的问题,但是没有意义,因为它没有 ROS 那样的广泛的代码库或社区。

I would never recommend writing your own simulator unless you are forced to. It is not a great learning experience, and it will never work as well as the ones that have been improved over many years by thousands of people.

With that said, the ones I have had positive experiences with are Gazebo and open dynamics engine (if you are using Matlab for your simulation). I would definitely recommend using Gazebo in combination with ROS. They are both becoming more standard and constantly improving. There are pre-existing simulations that you can base yours on like the erratic robot (for two wheeled differential drive platforms) and the PR2 (for humanoid type platforms). If you just want to do collision checking and kinematics, then I would recommend using rviz and the urdf format in ROS. Very easy to get started with and very well tested.

By the way, I have used Microsoft Robotics Developers Studio in the past, and it is not as good as ROS. I know they have a new version that fixes a lot of their previous issues, but there is no point because it does not have the extensive codebase or community that ROS does.

狼亦尘 2024-08-23 05:33:10

如果你熟悉C++编程语言,WEBOTS、ROBOWORKS、V-REP至少是三个兼容C++的仿真软件。完整的机器人仿真软件列表位于此处< /a>

If you're familiar with C++ programming language, WEBOTS, ROBOWORKS, V-REP are at least three simulation software C++ compatible. A comprehensive list with robotic simulation software is here

败给现实 2024-08-23 05:33:10

我最近一直在玩SimSpark。这是用于 RoboCup 3D 模拟足球联赛的平台。


(来源:google.com

I've been playing with SimSpark recently. It's the platform that's used for the RoboCup 3D Simulated Soccer League.


(source: google.com)

~没有更多了~
我们使用 Cookies 和其他技术来定制您的体验包括您的登录状态等。通过阅读我们的 隐私政策 了解更多相关信息。 单击 接受 或继续使用网站,即表示您同意使用 Cookies 和您的相关数据。
原文