CN810: Topics in Adaptive Mobile Robotics - SPRING 2014


PrerequisitesConsent of the instructor, Massimiliano Versace

Classes: Mondays, 11:00AM to 12:30PM, Room 302, 677 Beacon St. 

Lab: Tuesday, 4:00PM to 8:00PM, Room 308, 677 Beacon St.

Office hours: Wednesdays, 2:00 to 4:00 PM, and by appointment, Room 312, 677 Beacon St.

Description:  This course, first offered in 2010, is a hands-on introduction to the challenges of implementing adaptive behavior on a robotic platform (e.g., the iRobot Create, or a Romo). Using a simple hardware platform, students will design and implement computational models that allow the robot to perform behavioral functions, such as learning to recognize approach or avoid objects and locations in its environment (e.g., see here). 

This is not an introductory class in robotics.

Students are expected to have a sufficient interdisciplinary grounding in programming

 and the fundamentals of computational modeling of

mammalian vision, planning, and navigation to read primary research sources. 

Students will have the chance to use specialized neural software simulators (e.g., see here), virtual environments (e.g., see here), and the ASIMOV robotic interface. This will enable students to rapidly prototype large scale (millions of neurons, billions of synapses) on high-performance computing resources, testing them in realistic virtual environments. The 2014 edition will include at least one visit to a local robotic company (in the past, the class has visited iRobot and Kiva systems). 

Grades: they will be based on all aspects of a student's contribution to the course, including their oral presentations (project plan, progress on project plan, and demo days), their participation in class discussions, their ability to work in a team, and their course project. Students will be required to work in small groups. Students can choose any model to implement adaptive behavior on the elected platform, with the proviso that a "reality check" will be performed by the instructor on all projects to ensure that they are feasible and worthwhile. CN810 provides a great deal of independence in the choice of the particular robot, model and behavior focus of your final project. I expect that the final project would be both interesting to the student/team, and challenging, contributing eventually either to a publication or a dissertation prospectus, or anyway it will be worth remembering because it was fun and useful. 

Grades breakdown:

·         Project Plan (abstract and presentation, 20%)

·         Project Progress (20%, two 10% presentations)

·         Final Project (paper and demo/presentation/optional blog post, 50%)

·         Class Participation (including attendance, working in groups, etc, 10%).  

Timetable of deliverables: students taking the course for credit will submit a written research paper, present an oral summary of their research, and demo their model on the robot in the last weeks of the semester (demo days 1 and 2). Students are also required to submit the simulation code that is used to produce the demo. No interim reports are due, but three Power Point presentations are due on the project plan and progress (and issues…) on project plan days. Additionally, a brief abstract of the proposed project plan is due at the time of the project plan presentation. The third presentation, approximately 40 minutes long, will include demos of the robot, and occur on the demo days, as indicated below. Papers are due in paper and electronic format at least three days in advance of the demo day (this vary depending of whether you are scheduled to demo on day 1 or 2). An optional blog post can be submitted, as it has been done in the past (see here and here). A non-trivial portion of your final course grade will depend on the degree of professionalism that you exhibit in turning in that first draft in polished form and on time. 

Monday, March 3, 11:00 am: Project plan presentation (30 minutes per group). Turn in during class a hard copy and electronic copy of the abstract and of the slides for the proposal, consisting of a concise, informative title, a brief outline of your topic, including proposed methods and goals (20% of your grade).

Monday, March 31, 11:00 am: First progress presentation (30 minutes per group). Turn in a hard copy and electronic copy of the slides (10% of your grade).

 Monday, April 14, 11:00 am: Second progress presentation (30 minutes per group). Turn in a hard copy and electronic copy of the slides (10% of your grade).

 Monday, April 28, 5:00 PM, or Monday, May 5, 10:00AM, a copy of the presentation that you will give in the demo day is due via electronic submission, along with the final report, for group 1 and 2, respectively: Class presentation (45 minutes per group). You are required to turn in a complete and polished draft of your talk and report in Power Point, Adobe pdf, MS Word or some comparable format (no.... I do not use LaTeX, so if you use it, send a PDF). Failure to turn in a complete draft on time will be penalized. Your course grade will depend in part on the quality and completeness of your draft, as well as on the oral presentation itself. If your presentation includes animations, either transmit them as separate files or make them available on a web page or the wiki or for review (50% of your grade).

 Weekly Schedule – Classes: Mondays, 11:00AM to 12:30PM, Room 302, 677 Beacon St. Lab is on Tuesday, 4:30PM to 8:30PM, Room 308, 677 Beacon St. 




January 21


January 27

Class introduction

January 28

Lab: 1st part: Past projects accomplishments, calibrating 2014 goals, presentation of two ongoing robotic projects. Introduction to robot hacking. Byron Galbraith, Jeremy Wurbs, 2nd part: Robot hacking: Byron Galbraith, Jeremy Wurbs, Tim Seemann

February 3

Navigation using “place cells”. By Anatoli  Gorchetchnikov 

February 4


February 10

Modeling object recognition. By Florian Raudies

February 11


February 18


February 19

(Wed, Mon schedule)

Visual scene segmentation. By Florian Raudies

February 24

Visit to iRobot

February 25


March 3

Student presentations: present project plan

March 4


March 17

Reaching and grasping. By Byron Galbraith 

March 18


March 24

Collision avoidance in UAVs. By Jeremy Wurbs 

March 25


March 31

Student presentation: progress

April 1


April 7

Optic flow for mobile robots. Tim Barnes

April 8


April 14

Student presentation: progress

April 15


April 24

(Thu, Mon schedule)

Space-variant visual modeling. By Gennady Livitz 

April 28

Demo day 1

April 29


May 5

Demo day 2