Scalable and low-cost remote lab platforms: Teaching industrial robotics using open-source tools and understanding its social implications
Amit Kumar, Jaison Jose, Archit Jain, Siddharth Kulkarni, Kavi Arya · 2024
Read Scalable and low-cost remote lab platforms: Teaching industrial robotics using open-source tools and understanding its social implications by Amit Kumar, Jaison Jose, Archit Jain, Siddharth Kulkarni, Kavi Arya (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-15369.txt.
Introduction
Amit Kumar and co-authors describe, in a 2024 education paper, two cheap remote lab platforms that let students run real industrial robots from afar. Robot teaching is held back by the cost of the machines, safety risks, and hard training material. Their platforms rest on open-source robot software in two versions.
One setup runs a robot arm in a greenhouse as a farm harvest helper. The other runs a mobile rover in a warehouse system. Over seven months, more than fourteen hundred and thirteen hundred students used the two rigs, logging hundreds of remote control hours. The paper also weighs the social effects of spreading robot skills this way.
Catalog record
| CREATOR | Amit Kumar, Jaison Jose, Archit Jain, Siddharth Kulkarni, Kavi Arya |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-15369
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Scalable and low-cost remote lab platforms: Teaching industrial robotics using open-source tools and understanding its social implications Thanks: Accepted at Springer’s 16th International Conference on Social Robotics + AI 2024 Amit Kumar, Jaison Jose, Archit Jain, Siddharth Kulkarni, Kavi Arya Abstract With recent advancements in industrial robots, educating students in new technologies and preparing them for the future is imperative. However, access to industrial robots for teaching poses challenges, such as the high cost of acquiring these robots, the safety of the operator and the robot, and complicated training material. This paper proposes two low-cost platforms built using open-source tools like Robot Operating System (ROS) and its latest version ROS 2 to help students learn and test algorithms on remotely connected industrial robots. Universal Robotics (UR5) arm and a custom mobile rover were deployed in different life-size testbeds, a greenhouse, and a warehouse to create an Autonomous Agricultural Harvester System (AAHS) and an Autonomous Warehouse Management System (AWMS). These platforms were deployed for a period of 7 months and were tested for their efficacy with 1,433 and 1,312 students, respectively. The hardware used in AAHS and AWMS was controlled remotely for 160 and 355 hours, respectively, by students over a period of 3 months. Keywords: Educational robots; Robot Operating System (ROS); ROS 2; Industrial robots; Agricultural robots 1 Introduction and Related Work One of the key aspects of Industry 5.0 is to adopt a human-centric approach to digital technologies [ 3 ] . With the world transitioning to this fifth industrial revolution, training students in current and upcoming technologies is necessary. However, access to the latest technologies, like those used in industrial robotics, is expensive. As a result, only a handful of students tend to learn them. Most manufacturers of industrial robots provide a model of their robots that can be imported into various simulators [ 22 , 32 ] . However, in general cases, a sim-to-real [ 12 ] gap is prevalent. Hence, it is a well-adapted practice to develop rudimentary algorithms in simulation and then make them more robust by testing them in hardware. ROS [ 24 ] and ROS 2 [ 19 ] provide a robust modular platform for easy and fast deployment of robotic applications. They are open-source; hence, many existing libraries, in the form of ROS packages, can be configured for a particular application instead of being rewritten. Many industries and organizations like NASA [ 23 ] have also harnessed the power of ROS in their projects. It is then only logical to train students in industrial robotics using ROS. Figure 1 : Set up of remote testbeds in a greenhouse (top) and a warehouse (bottom) for Autonomous Agricultural Harvester System (AAHS) and Autonomous Warehouse Management System (AWMS) using a UR5 robotic arm and a mobile rover. Students used our platforms to test their algorithms in the simulated-world (left) and real-world (right). Setting up designated laboratories to provide remote access to industrial robots is not unheard of [ 2 , 7 , 10 , 17 ] . These remote labs bridge the gap between buying multiple expensive robots and giving access to many students. Several works have proposed and tested different architectures for teaching students through remote robotic labs. Researchers in [ 16 ] have created a web app and integrated virtual network computing (VNC)-based graphical desktop-sharing and docker containers. The web app allows the users to access the small Robotont mobile robot and the UFACTORY xArm7 manipulator robot individually. They have deployed their robots in a simple and small arena, which may not reproduce all the complications generally accompanied while deploying industrial robots in real-world settings like noise in sensors due to direct light, heat, etc. A low-cost and lightweight robotic manipulator using Raspberry Pi 3 computer is deployed in [ 33 ] to teach primary-school and university-level students. Although the cost of the robotic arm is low, these arms do not offer industrial standard functionalities like force-torque sensing. The payload and reach of these arms are also unsuitable to be deployed in a real-world setting. ROS was not integrated into their architecture since they aimed to introduce basic manipulation concepts to young students. Researchers in [ 2 ] transfer the script to the NAO humanoid robot initially and then run the script locally. However, the user cannot obtain live feedback during the execution of the code. KUKA Robot Learning Lab [ 31 ] uses a queue-based parallel processing architecture for running code onto static robotic arms. Whenever a user submits a code using their web interface, it is first tested in a simulator, Gazebo, and then sent to the actual robot. Although this improves the safety of the hardware, it comes with a computational overhead. Moreover, in a real-world setting, as shown in AAHS and AWMS, where a depth camera, robotic arm, and a mobile robot work in parallel, this can significantly slow the execution. Integrating the safety of the robots in the hardware itself may be better suited for scalable applications. Additionally, none of the remote robotic labs mentioned above have tested their architecture on a large scale. These methods may not function as intended or even be expensive when thousands of students use them. In totality, developing efficient platforms, encompassing simulations, for remote labs using open-source tools like ROS is an important and scalable step to prepare our students for Industry 5.0. One of the learning-based methodologies, project-based learning (PBL), has shown to be an effective method for learners to master new concepts, including ROS [ 8 , 11 , 15 , 20 , 21 , 27 ] . Our research focuses on developing these remote lab platforms and testing their efficacy by deploying them in our PBL-based competition called the e-Yantra Robotics Competition (eYRC). Our platforms were used by students to develop algorithms for an Autonomous Agricultural Harvester System (AAHS) and Autonomous Warehouse Management System (AWMS), as shown in Figure 1 . These platforms helped students learn about image processing, navigation algorithms for mobile rovers, motion planning algorithms for manipulators, Proportional Integral Derivative (PID) controllers, and ROS. The major contributions of this work are as follows: 1. Design of two low-cost, scalable, and customizable platforms using open-source tools for teaching industrial robotics remotely. 2. Study our platforms’ effectiveness by deploying them in a PBL-based competition spread over six months with more than 1000 students. 2 Methodology 2.1 Setup of Platforms Scalability is one of the main factors considered while designing our platforms. While adding multiple work cells can easily increase the number of students that access a remote laboratory, it comes at an additional cost. Even though the simulation created, as shown in Figure 1 , for each platform closely mimicked the real-world, we can not eliminate the sim-to-real gap. Hence, in both of our platforms, students first tested their algorithm using a simulator, Gazebo [ 14 ] , and then only selected students (50-100), based on their performance, tested them using remote hardware. The students connected to the hardware in one-hour slots for three months. These students participated by forming a team of 2-4 members in our 7-month eYRC competition. The multi-robot system (MRS) used in both of our platforms is shown in Figure 1 . The four-wheeled mobile