Efﬁcient delivery of Robotics Programming educational content using Cloud Robotics Sean Murphy Leonardo Militano Giovanni Toffetti and Remo Maurer

2025-05-03 0 0 2.77MB 6 页 10玖币

侵权投诉

Efﬁcient delivery of Robotics Programming

educational content using Cloud Robotics

Sean Murphy, Leonardo Militano, Giovanni Toffetti and Remo Maurer

Zurich University of Applied Sciences

Switzerland

Email: [murp |milt |tof f |murm]@zhaw.ch

Abstract—In this paper, we report on our use of cloud-robotics

solutions to teach a Robotics Applications Programming course

at Zurich University of Applied Sciences (ZHAW). The usage of

Kubernetes based cloud computing environment combined with

real robots – turtlebots and Niryo arms – allowed us to: 1) min-

imize the set up times required to provide a Robotic Operating

System (ROS) simulation and development environment to all

students independently of their laptop architecture and OS; 2)

provide a seamless “simulation to real” experience preserving

the exciting experience of writing software interacting with the

physical world; and 3) sharing GPUs across multiple student

groups, thus using resources efﬁciently.

We describe our requirements, solution design, experience

working with the solution in the educational context and areas

where it can be further improved. This may be of interest to

other educators who may want to replicate our experience.

Index Terms—Cloud Robotics, Kubernetes, Robotic Applica-

tions, Edge Computing.

I. INTRODUCTION

The School of Engineering at ZHAW has been offering

the “Robotic Applications Programming” (RAP) course to

bachelor students since 2021. The course is intended for IT

Bachelor students as a way to 1) learn how to program ap-

plications using ROS-based robots, and 2) leverage interdisci-

plinary knowledge acquired during the study programme (e.g.,

Artiﬁcial Intelligence, computer vision, distributed systems,

cloud, Operating Systems, web- and mobile-development) and

integrate them to achieve autonomous robotic behavior. We

focus on ROS as it is currently the most used framework,

it is Open Source, and has ever-increasing capabilities with

many contributions of advanced algorithm implementations

and robotic simulation packages [5], [4].

The course is organized in three main sections as follows:

(1) students ﬁrst learn ROS fundamentals (communication

primitives and building ROS packages) and robotics (e.g.,

basic robotic Hardware, robot models and visualization, coor-

dinate frames and transformations, controllers); (2) additional

base capabilities are learned (e.g., Simultaneous Localization

and Mapping - SLAM, navigation, perception, arm motion

planning and control); and (3) ﬁnally these are combined to

build a practical application for the yearly challenge. We do

not explicitly consider mechanical engineering aspects nor

system integration aspects that complete the robotics engi-

neering ﬁeld. This year’s challenge is inspired by the DARPA

Subterranean Challenge1: students will have to write software

to control an autonomous mobile manipulator – a simulated

Summit XL2with a UR-5 arm3– in an unknown environment

performing mapping, pose estimation and collection of known

objects and returning all objects to the starting location.

During the course of the semester, students apply the

theoretical concepts they learn in class to lab sessions; the

earlier lab sessions use simulated robots for quick software

development cycles and the later lab sessions run the same

software to control real robots - 6 turtlebot3’s are used for

SLAM/navigation and 3 Niryo arms are used for grasping.

In order for the students to concentrate on course content

and minimize the time they would need for system set up and

conﬁguration, we needed to prepare some teaching infrastruc-

ture. We had the following key requirements:

•R1: Provide a consistent collaborative environment for

group work across multiple access devices (tablets, lap-

tops with different Operating Systems and CPU architec-

tures);

•R2: Support Simulation with a realistic simulated-to-real

time ratio and frame-rate;

•R3: Support transitioning from the simulated environ-

ment to real world robots with minimal effort

The main contribution of this paper then is the system

design which meets these requirements. Possible technologies

are described, evaluated and the design choices for the ﬁnal

solution are discussed.

The paper is structured as follows. In section II, we review

related work broadly classifying solutions into simulation

focused solutions, hardware focused solutions and hybrid

solutions. Section III describes our solution including the

basic components and how they ﬁt together. In section IV

we discuss our experience using the platform in the classroom

environment. Section V discusses open issues with the current

solution and ﬁnally there is a conclusion in section VI.

II. RELATED WORK

The interest in robotics engineering has been growing

rapidly over the last few years. For hobbyists, students and

professionals, the amount of robotics practitioners has steadily

1https://www.subtchallenge.com/

2https://robotnik.eu/products/mobile-robots/summit-xl-en-2/

3https://www.universal-robots.com/products/ur5-robot/

arXiv:2210.10441v1 [cs.RO] 19 Oct 2022

grown and with it the available educational content. At the

same time, educational institutions at all levels are working

hard to adapt their instructional programs and learning paths

to integrate robotic technologies. For instance, Educational

Robotics (ER) is a modern teaching practice that the teacher

engages in, using robots as a tool for designing and integrating

the educational process. ER was identiﬁed as an educational

resource through which students acquire knowledge of dif-

ferent disciplines and improve their attitude and interest in

STEAM disciplines (Science, Technology, Engineering, Arts

and Mathematics) [1], [3].

Depending on the educational level and the requirements for

professional knowledge of robotic application development,

different teaching and learning approaches can be identiﬁed.

Here, we categorize them as follows: i) simulation-based;

ii) hardware-based; and iii) combination of simulation- and

hardware-based solutions.

Simulation-based learning leverage software tools and

programming languages to simulate the behavior of robots

without direct interaction with a physical robot. Under this

category we include web robotics as a way of learning online

using a web-based platforms for simulating robots, as e.g. in

[2]. This latter is gaining momentum with offerings such as

AWS RoboMaker4which are cloud-based simulation services

that enable robotics developers to run, scale, and automate

simulation without managing any infrastructure. One of the

most widely used simulators for ROS is Gazebo5which

provides 3D physics simulation for a large variety of sensors

and robots. Gazebo is bundled into the full installation package

of ROS, making it widely and easily available, and many

robot manufacturers offer ROS packages speciﬁcally designed

to support Gazebo. Other popular robotic simulators are We-

bots6, CoppeliaSim7and OpenRave8. Besides these, game

engines are also being adapted to support robotic simulation

such as, for instance, Unity9. Simulation based solutions are

clearly useful and serve some important educational needs;

however, the models on which they are based always have

some limitations which can become apparent in a real world

context. Further, adopting a simulation only approach does

not give students experience with some of the more practical

considerations associated with working with physical devices.

Hardware-based learning focuses on direct interaction and

programming of physical robots. In some simple domains and

for simple applications students can safely interact directly

with the hardware without necessarily having ﬁrst simulated

the application behavior. One example of this is the LEGO®

Robot Programming for kids program10 where kids build

a robot, program it and interact with it; programming in

this environment is based on a set of predeﬁned tasks the

4https://aws.amazon.com/robomaker/

5https://gazebosim.org/

6https://cyberbotics.com/

7http://www.coppeliarobotics.com/

8http://www.osrobotics.org/osr/

9https://github.com/Unity-Technologies/Unity-Robotics-Hub

10https://www.lego.com/en-gb/categories/coding-for-kids

robot can execute. Similar solutions based on compositions of

predeﬁned tasks resulting in more complex behaviours exist

(e.g., the Blockly interface of the Niryo Ned robotic arm11).

Such solutions, however, lack ﬂexibility and the extensibility

and customization capabilities required for real world robotics

scenarios. To develop more realistic applications the use of

programming languages such as Python, C++, MATLAB or

ROS is a must. Moreover, in complex environments, where

access to hardware is not always possible or too expensive, it

becomes also mandatory to test the application behavior in a

simulated environment ﬁrst.

Hybrid learning combining simulation and hardware-

based learning is a solution in which the robotic applica-

tion can be tested in a simulated environment and deployed

on the physical devices in either a two-steps process or

in hybrid manner. In the two-steps process where we keep

simulations (ﬁrst step) separate from testing on real hard-

ware (second step). In doing so we have the advantages of

less costs, reduced risks of damaging expensive hardware,

reduced risks of damages to third persons and things. In a

hybrid approach, concepts like digital-twin gain importance

for developing robotic applications/tasks. A digital copy of a

robotic hardware can be used for visualization and control of

the robot. In advanced solution, a digital-twin can be placed

into a simulated environment while the actions and tasks are

physically executed on the hardware. In this way, the simulated

environment will provide inputs to the application in terms of

environment (e.g., obstacles), sensing information (e.g., light,

temperature), which allows to test applications in a close-to-

real environment.

As the complexity of robotic applications is growing

steadily, with the adoption of advanced analytic solutions such

as Artiﬁcial Intelligence, Semantic Navigation, Autonomous

motion, new needs appeared in terms of computation, network-

ing and storage resources. To cope with them, Cloud-based

solutions started to see the light for computation ofﬂoading

on the Cloud. The possibility for remotely controlling robotic

systems further reduces costs for deployment, monitoring,

diagnostic and orchestration of any robotic application. This,

in turn, allows for building lightweight, low cost and smarter

robots as the main computation and communication burden is

brought to the cloud. Since 2010, when the Cloud Robotics

term ﬁrst appeared, a number of projects (e.g., RoboEarth

[8] DAVinci [9]) investigated the ﬁeld pushing forward both

research and products to appear on the market. Companies

started investing in the ﬁeld as they recognized the huge

potential of cloud robotics. This lead to ﬁrst open source cloud

robotics frameworks appearing in recent years. An example

of these is the solution from Rapyuta Robotics12. Similarly,

commercial solutions for developers have seen the light with

the big players in the Cloud ﬁeld joining the run (e.g., Amazon

Robomaker and the Google Cloud Robotics Platform13).

11https://niryo.com/robotic-solution-education-research/

12https://www.rapyuta-robotics.com/

13https://cloud.google.com/cloud-robotics/

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

EfcientdeliveryofRoboticsProgrammingeducationalcontentusingCloudRoboticsSeanMurphy,LeonardoMilitano,GiovanniToffettiandRemoMaurerZurichUniversityofAppliedSciencesSwitzerlandEmail:[murpjmiltjtoffjmurm]@zhaw.chAbstractInthispaper,wereportonouruseofcloud-roboticssolutionstoteachaRoboticsApplicationsP...

展开>> 收起<<

Efﬁcient delivery of Robotics Programming educational content using Cloud Robotics Sean Murphy Leonardo Militano Giovanni Toffetti and Remo Maurer.pdf

共6页,预览2页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Efﬁcient delivery of Robotics Programming educational content using Cloud Robotics Sean Murphy Leonardo Militano Giovanni Toffetti and Remo Maurer

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: