Thursday, December 14, 2017
Home » EduBots » Fostering Innovation through Robotics Exploration (FIRE)

Fostering Innovation through Robotics Exploration (FIRE)

Robot magazine interviewed key associates working at the Carnegie Mellon Robotics Academy to get an update on FIRE programs now in development. Heres a snapshot of work that is proceeding with respect to MultiRobots (collaborative swarmbots), RobotC expansion to include the Arduino controller and simulator software, Cognitive Tutor technology and what is coming in FIRE Virtual Worlds that will enable students to program and test robots online. Be sure and visit http://fire.cmu.edu for the very latest FIRE news!

This photo was taken during a design review in which the projects technical development team conducted a demonstration for the curriculum development team. Platforms shown are Arduino Uno, Innovation Firsts VEX Cortex, LEGO MINDSTORMS NXT, and the TETRIX add-on kit for the NXT.
MultiRobot UpdateCarnegie Mellons Robotics Academy is beginning to move into another dimension in educational robotic technology. Early initiatives focused on building remote control robots, which then evolved into programming autonomous robots with various platforms. The next stage of educational robotics will involve merging multiple types of autonomous robots to communicate together and share decision-making and tasks to solve complex challenges. The project team has multiple robots controllers talking to each other using Xbee radios. The robot platforms that the team is developing with are the Arduino Uno, Innovation Firsts VEX Cortex, LEGO MINDSTORMS NXT, and the TETRIX add-on kit for the NXT.Carnegie Mellons FIRE project is developing and iteratively testing an inexpensive educational robotics platform that can be used in schools that uses common off the shelf (COTS) robotics technology capable of multi-robot communications. Carnegie Mellon has integrated the Xbee communications protocol into their popular multiple-platform robotics programming language, ROBOTC, allowing ease of use when communicating with multiple robots.Another part of the project will develop a set of scaffolded training materials (challenges, lessons, and web support) enabling teachers and coaches to teach students how autonomous multi-agent communications works; examples include: robot search and rescue, robot soccer, robot dancing, and automated work cells. Carnegie Mellon plans on integrating this technology into multiple competitions within the next two years. FIRST Tech Challenge, Robocup, and VEX Robotics Competition have all expressed interest in the multi-robot project.
–Timothy Friez
Carnegie Mellons ROBOTC Targets the Arduino The folks at Carnegie Mellon are porting their popular educational multi-platform robotics programming language, ROBOTC, to the Arduino series of microcontrollers. The Arduino version of ROBOTC will support the new Uno series of controllers as well as the older Duemilanove (ATMega328) and Mega (ATMega1280) boards. ROBOTC will bring exciting new features to the Arduino, such as interactive debugging capabilities without the use of expensive additional hardware, multiple motor shield support and compatibility with existing motors and sensors from VEX Robotics. Users will also be able to take advantage of XBee wireless radios adding wireless programming and multi-robot communications to their Arduino boards with ROBOTC.Carnegie Mellons Robotics Academy will also be developing tutorials and lessons to help teachers and students get up and running with programming and projects. A Beta version of the software will be available this winter at the ROBOTC website. Read more about using ROBOTC with the Arduino, VEX, or LEGO platforms and all of the latest news by visiting www.robotc.net.
–Timothy Friez


ROBOTC Simulator Software

The Robotics Academy at Carnegie Mellon University is developing a robot simulator for the popular ROBOTC language. Students, hobbyists and adults will be able to learn to program a virtual robot using the same language and programs as other robots supported by ROBOTC. The virtual environment of the simulator will contain state of the art 3D graphics, real time physics, and video game-like environments. Users will have a choice of popular robots with sensors such as touch sensors, encoders, magnetic compasses, and sonar distance sensors. More information, videos, and screen shots are available at http://fire.cmu.edu by following the Virtual Worlds link.
–Ed Paradis

Cognitive Tutor Enabled Robotics Training

Robots are invading our schools. Small, inexpensive, customizable robot systems such as the LEGO® MINDSTORMS® NXT and VEX Robotics systems have begun making headway into classrooms and afterschool programs across the world. The popularity of these school-friendly technologies opens the doors to a hands-on revolution in CS-STEM (Computer Science, Science, Technology, Engineering, and Mathematics) education. And yet, preliminary research shows that the academic elements especially math are often left out in practice (see http://robotics-academy.org/blog/?p=356).

Carnegie Mellons Robotics Academy has teamed up with the Cognitive Tutor Authoring Tools project to solve that problem by bringing AI-enabled automated tutoring to the classroom. A Cognitive Tutor (CT) is a simple artificial intelligence programmed to recognize certain common mistakes students might make as they work through problems online, and offer guidance to correct them.

The first CT-enabled robotics curriculum module under development, Robots in Motion, takes students through several commonly-encountered Robotics situations in ways that deliberately encourage students to use, rather than avoid, mathematics as a problem-solving tool. Abstract ideas about ratios, rates, angles, and unit conversions suddenly become concrete situations involving distances, speeds, turns, and measurement.

As a student works through the lessons, the AI Tutor monitors his or her progress and steps in whenever appropriate to provide individualized guidance, whether the problem is conceptual or technical in nature. Students are thus more likely to complete the lesson successfully, and teachers are free to spend more time helping students with big questions by leaving the little mistakes for the AI to correct.

The current means of communicating with the Cognitive Tutor is through the on-screen dialogs and widgets. The lesson, however, is constructed to work with (and rely on) concurrent robot use. The rightmost column in the measurement and prediction table, for instance, asks students to run the robot using the given rotation settings to test whether their mathematical model really made good predictions or not.The Academy is planning to extend the tutorable environments far beyond the existing 2D widget set. The tutor model could theoretically track the physical parameters a student chooses for a virtual robot. It could even monitor and give feedback about the code that a student types into ROBOTC for an assignment. This sort of tutoring has already been done by other researchers using Intelligent Tutoring Systems. At some point in the foreseeable future, the Tutor may be catching infinite loops, backward > and < signs, and missing semicolons long before they ever require a teacher’s attention.

At the end of the day, the teacher can check on each students progress using a built-in set of teacher tools. The teacher can review individual results, or generate high-level reports that reveal both students who need additional attention, and topics which may need extra explanation. Combining a robotics curriculum that embraces CS-STEM with a built-in AI teachers assistant will allow robotics clubs and classrooms worldwide to provide learning experiences for students that are both valuable and manageable.The Robotics Cognitive Tutor Project is one of several projects being developed under a CS-STEM Education grant from DARPA. Together, these projects collectively known as FIRE (Fostering Innovation through Robotics Explorations) seek to increase interest and improve the quality of understanding that future generations have in Computer Science and STEM-related disciplines. Want to learn more? Check out a free preview and sign up for updates at http://fire.cmu.edu.
–Ross Higashi and Vu Nguyen


To learn more about the Fostering Innovation through Robotics
Exploration (FIRE) program, Click Here