Teaching robotics everywhere

Abstract

The Robotics Education Workshop (http://projects. csail.mit.edu/rss/RobotEd/) took place in connection with the 2005 Robotics Systems and Science (RSS) symposium (http://www.roboticsconference.org/) at the Massachusetts Institute of Technology (MIT) on 11 June 2005. It was organized by John Leonard, Una-May O'Reilly, Nick Roy, Daniela Rus, and Seth Teller. The workshop’s main goal was to discuss what it would take to turn robotics into a core course that can be taught in every accredited CS, ME, and EE undergraduate program in the United States and elsewhere around the world. We were inspired to organize this workshop during an introductory robotics course we have been developing as a team (http://courses.csail. mit.edu/6.141) and were encouraged by Hal Abelson. There are many reasons why we believe now is an important time to consider teaching robotics broadly. The computing revolution is bringing computers closer and closer to interactions with the physical world. We have seen the progression from mainframes to desktops to laptops to pervasive computing by networking. Most of today’s computers have integrated sensors, such as microphones, primitive motion sensors to park the disk drive, light sensors to detect when to automatically illuminate the keyboard or dim the screen, and cameras to collect images from the surrounding environment. Recent advances in the cost, size, performance, and capabilities of actuators show us that actuation will become native to a computing environment so that today’s personal computers (PCs) will become tomorrow’s personal robots (PRs). Robotics brings an interesting perspective to computer science education as it naturally puts together continuous and discrete computation and provides an opportunity to reason about error and uncertainty, an important new topic in the study of computation at the undergraduate level. Last but not least, the love of robots is as pervasive as can be. Robotics is hard to teach because the expertise in robotics is concentrated at a few institutions. And, there are no established curricula, no robust and inexpensive teaching platforms, and no collections of labs and projects with solutions. Teaching robotics well relies on integrating together ME, EE, and CS topics; creating a balanced syllabus is a huge challenge. Yet, robotics provides the perfect educational tool for introducing students to embedded systems and computation for interacting with the physical world, and to making intelligent autonomous machines. Several universities have already introduced special topics courses on robotics. The curricula and hardware platforms for these courses are quite varied, and the goals of the courses differ across ME, EE, and CS departments. We wished to leverage this excellent body of knowledge and discuss how to develop an integrated approach to teaching robotics that would train students simultaneously in foundational aspects of designing, controlling, and programming robots and embedded systems. To this end, the discussions at the RSS education workshop were focused on 1) evaluating the state of the art for undergraduate robotics education, 2) discussing how to build on this experience toward a broad integration of robotics in the undergraduate curricula, and 3) discussing how to best share course materials. Over 30 professors from universities and colleges in the United States, Europe, and Asia met for a full day. The program alternated presentations on existing courses and educational material with lively discussions on the key issues of integrating robotics in an undergraduate curriculum. The workshop program included presentations by Gaurav Sukhatme (University of Southern California), Mark Yim (University of Pennsylvania), Illah Nourbakhsh (Carnegie Mellon University), Brad Nelson (ETH-Zurich), Jonathan Bredin (Colorado College), Holly Yanko (University of Massachusetts, Lowell), Matthew Stein (Roger Williams University), Matt Mason (Carnegie Mellon University), John Hollerbach (University of Utah), Nick Roy (MIT), and Stewart Tansley (Microsoft). Three discussion sessions focused on 1) hardware kits, software packages, and lessons learned (moderated by Una-May O'Reilly, MIT); 2) integrated syllabus and projects (moderated by Seth Teller, MIT); and 3) what to share and how (moderated by Daniela Rus, MIT). The presentations were all exciting, inspiring, and filled with interesting lessons learned, and workshop participants were highly engaged in the discussion sessions. The most important outcome was a general goal of creating an open repository of robotics course materials in order to better coordinate the teaching of robotics across universities. Such a repository would enable the sharing of course materials. By providing easy access to teaching materials, the teaching of introductory robotics courses everywhere will be facilitated, and robotics will become a core component of computer science and engineering academic programs. This will strengthen the education and research activities for our field. More specifically, the participants discussed the possibility of developing an open repository of robotics course materials similar to MIT's OpenCourseWare (OCW) initiative. This repository would contain a suite of robotics curricula for teaching all aspects of robotics, ranging from undergraduate introductory courses E D U C A T I O N