M&AE 4180/5180, CS 3758Autonomous Mobile RobotsSpring 2012Creating robots capable of performing complex tasks autonomously requires one to address a variety of dif-ferent challenges such as sensing, perception, control, planning and interaction with humans. Many advanceshave been made in recent years towards creating such systems, both in the research community (differentrobot challenges and competitions) and in industry (industrial, military and domestic robots).This course will give an overview of the challenges faced and techniques used for creating autonomous mobilerobots. Topics include: sensing, localization, mapping, path planning, motion planning, obstacle and collisionavoidance and multi-robot control. The course includes a lab portion in which students program the iRobotCreate.Note: This course is very hands-on, requiring students to implement algorithms, debugcode and experiment with physical robots. As such, students should have a good workingknowledge of Matlab, and should start assignments early.Instructor: Prof. Hadas Kress-Gazit email:[email protected], office: 210 Upson HallTA: Ben Johnson email:[email protected]: Kevin Wyffels email:[email protected] and Location:Lectures: Tue. and Thur. 10:10 - 11:25 203 PhillipsLab sections: Wednesday 10:10AM - 12:35PM 102A ThurstonThursday 2:00PM - 4:25PM 102A ThurstonFriday 2:00PM - 4:25PM 102A ThurstonThere will be five labs throughout the semester.Prof. Kress-Gazit’s office Hours: Wed 1pm-2pm, 210 Upson Hall or by appointment.TA office Hours: Mon 4:30-5:30PM and Thur 5-6PM in 102A Thurston.Course website: http://web.mae.cornell.edu/hadaskg/courses/mae4180.htmAnnouncements, assignment posting, submission and grading will be done through the Course ManagementSystem (CMS). Students log in using netid to http://cms.csuglab.cornell.edu.References:Course Textbooks:• S. Thrun, W. Burgard, and D. Fox, Probabilistic Robotics, MIT Press, 2005. (TBF)• H. Choset, K. M. Lynch, S. Hutchinson, G. Kantor, W. Burgard, L. E. Kavraki and S. Thrun, Principlesof Robot Motion: Theory, Algorithms, and Implementations, MIT Press, Boston, 2005. (Choset)Additional sources:• R. Siegwart and I. R. Nourbakhsh, Introduction to Autonomous Mobile Robots, MIT Press, 2004.• S. M. LaValle, Planning Algorithms, Cambridge University Press, 2006.Available online at http://planning.cs.uiuc.edu/• M. J. Matari´c, The Robotics Primer, MIT Press, 2007.Workbook online at http://roboticsprimer.sourceforge.net/ workbook/Main PagePrerequisites: Junior, Senior or Graduate standing or permission of instructor. Good knowledge ofprogramming (MATLAB) is essential for the homework and labs.Laboratory componentThis course includes a laboratory component so that students can gain first-hand experience with thealgorithms discussed in the lectures. The labs are centered around the iRobot Create platform and involvecollecting data, programming the robot and discussing its behavior. All labs include a pre-lab assignmentthat will be part of the homework sets, a lab manual describing the experiment and a post-lab writeup.Reports are expected to be concise and to the point.Simulator: A MATLAB-based simulator for the iRobot Create will be used in this class in order to facilitatecode development for the labs and the final competition.The simulator can be found at http://web.mae.cornell.edu/hadaskg/CreateMATLABsimulator/createsimulator.htmlLaboratory Schedule:Week 03 (02/6): Lab 1 Introduction to Controlling the iRobot CreateWeek 06 (02/27): Lab 2 LocalizationWeek 08 (03/12): Lab 3 MappingWeek 11 (04/2): Lab 4 Potential Field Motion PlanningWeek 13 (04/16): Lab 5 Sampling-based Motion PlanningFinal Competition: Thursday, 5/3 Last Day of ClassCoding: To truly understand the concepts and algorithms discussed in the lecture, students will implementmost of the algorithms and experiment with simulated and physical robots. Good coding practices willfacilitate the implementation and debugging phase and will allow for intelligent code reuse. Students arehighly encouraged to be mindful of their code structure and consult the professor and the TAs about waysto improve reusability.Prof. Kress-Gazit will hold a coding recitation on Friday, Jan 27th, 2:30-3:30 PM in Upson 109to convey the principles of good coding practices.PoliciesLecture notes: Lectures will be presented using a tablet PC. A skeleton file for the lecture notes will beposted at least an hour before each lecture (typically the night before) and it will mainly include figures andsome text. It is recommended that students come to class with the lecture notes skeleton and use it to fillin the material covered in class. The lecture notes will be posted after the class but they will NOT containall the information and discussion that takes place in class.Labs:• The labs are an integral part of the course and as such, in order to receive a grade, students mustcomplete all of the labs.• Makeup labs for students who miss a lab will be scheduled on the week of May 7th. Students areexpected to notify the TAs if they are unable to attend a lab.• Lab groups will consist of 2-3 students. Groups will be assigned by the TAs and will change every lab.• Pre-lab assignments (code needed for the lab, part of the homework sets) will be done individuallyand the post-lab report will be written as a group (one report submitted per group).• To make the most out of the time in the lab, and to focus on working with the robots and not ondebugging code, all students must arrive at the lab with working code. Students will not beallowed to do the lab if they do not have working code needed for the lab. If a student wishes to usea late day, they must coordinate with the TAs to attend a different lab section.• During labs, all group members must remain in the lab until data is collected by all group members,unless approved by the TA.2• Post-lab reports will include a sentence describing the contribution of each team member to the analysisand report.Grading: The grade will be determined based on the following: pre-class assignments, homework sets,labs and a final competition.• Pre-class assignments: To facilitate the learning process and maximize the effectiveness of thelectures, Prof. Kress-Gazit will assign reading before each lecture. In addition, short problem sets(typically one question) will be posted two days before class and will be due (paper, not electronicsubmission) at the beginning of class. (10% of the final