COMP150

kramer-the-robot

Behavior-Based Robotics

Fall 2010

Syllabus

COMP150 "Behavior-Based Robotics" (3 cr)

This course is designed to investigate and study methods and models in embodied cognitive science and artificial intelligence, with particular focus on behavior-based techniques on robots.  All models and architectures will be theoretically scrutinized and evaluated with respect to their conceptual clarity, support by empirical data, plausibility, etc. without neglecting issues of practicality such as feasibility of implementation, real-time/real-world issues, computational resources, etc. These practical considerations will turn out to be particularly important for model implementations on robots.

Required Course Text:

Arkin, C. Ronald (1998), Behavior-Based Robotics, MIT Press, Cambridge: MA.

Additional (Optional) Readings:

Pfeiffer, R.. and  Scheier, Ch. (1999) Understanding Intelligence.  MIT Press, Cambridge: MA.Murphy, R. (2002) Introduction to AI Robotics. 2nd ed.  MIT Press, Cambridge: MA.
Bekey, G. (2005) Autonomous Robots: From Biological Inspiration to Implementation and Control (Intelligent Robotics and Autonomous Agents). MIT Press, Cambridge: MA.

In addition, various seminal research articles will be made available and read throughout the course.

Instructor: Matthias Scheutz

E-mail: mscheutz@cs.tufts.edu
Office: 107B Halligan Hall
Office Hours: by appointment only
Phone: (617) 627-0453 or internal 7-0453

Teaching Assistant: Jisoo Park

E-mail: jpark@cs.tufts.edu
Office: 107 Halligan Hall
Office Hours: TBD and by appointment

Course Goals:

There are three goals of this course: (1) to present an overview of the state-of-the-art in behavior-based robotics research and its implications for embodied cognitive science, incorporating results from various different subdisciplines in cognitive science (including artificial intelligence, robotics, ethology and psychology); (2) to give students an appreciation of how "hard" robotics is, because the real world (different from many simulation environments) is full of imperfections, noise, and failures that successful robot models have to cope with; and (3) to foster the students' ability to work in (interdisciplinary) groups on original complex software projects to tackle research questions that can lead to publishable results.

 Topics:  Appr. number of lectures 
 Introduction to embodied cognitive science and behavior-based robotics
4
 Overview of the ADE development environment and the available robots 
2
 Reactive behavior-based architectures
6
 Perception
4
 Deliberative systems
6
 Hybrid systems
2

Note that the number of lectures on each topic (as well as some of the various subtopics covered by the lectures) may vary depending on the respective research projects picked by students.

Computer Usage:

Programming in the course will be done mostly in JAVA (in the ADE environment) and possibly other languages depending on individual preferences for the group projects.

Grading:

The main goal of this course is to get robots to do something that elucidates an interesting principle, property, or function of cognition.  Hence, students will spend most of their time working in groups (consisting of three to four members) on a project of their choice.  They will start working on it as early as the third week of classes and present regular progress reports in class (each group member will have to present at least once).  At the end of the semester, groups will turn in a final paper written in the style of a conference contribution (ideally, this is what it should be used for!).  Their grade will reflect the quality of their attempts of getting their models to work, but will also depend on the performance on individual assignments (during the first few weeks of the semester), the (individual) final report and on class presentations.


Undergraduate students: Graduate students:
The following grade breakdown will be used:
 
 92 - 100  A 
 89 - 91  A-
 86 - 88   B+
 82 - 85  B
 79 - 81  B-
 76 - 78  C+
 72 - 75  C
 69 - 71  C-
 62 - 68  D
   0 - 61  F

Class presentations and group assignments

Each graduate student will present one research paper to the whole class in the course of the semester.  This presentation will be the main component of the additional graduate student effort. Before the group projects start, there will be a few structured lab assignments to get students up to speed on the employed robots and software development environments.  All assignments (including code) that have to be turned in will be submitted electronically by email to the TA.

Late Policy

Late assignments are in general not accepted, hence will not earn any credit (except in extraordinary formally documented circumstances).

Class Attendance and Participation

Although no rigorous attendance policy will be implemented for this course, students are expected to attend all classes (students with excessive absences will be very unlikely to pass the course).  Everybody is encouraged to participate actively and contribute to the course (e.g., by asking questions and sharing information in the online web-based forum).

This course follows the Faculty of Arts, Sciences and Engineering Guidelines Pertaining to Religious Observances. You are not required to prove attendance at religious services or events to obtain an accommodation for religous observance, but you are requested to provide indication of such any accommodation requests early in the semester.

Incompletes:

Per Tufts policy, incompletes will be granted under only the most exceptional of circumstances (out of your control) and only in cases where most of the course work has already been completed. Examples of exceptional circumstances include a death in the family or major illness that keeps you out of the classroom for a significant period of time. Getting behind in the class due to other obligations outside the classroom (other classes, job) doesn't warrant granting an incomplete (for details see the .

Final:

There will be a final research paper (instead of a final exam) due at the end of the course, written in the style of a conference publication and graded based on the criteria used for conference publications (guidelines and templates will be provided).

Academic Honesty:

This course is conducted in accordance with the Academic Integrity standards as described in the School of Arts and Sciences / School of Engineering Academic Integrity booklet.  Specifically, it is considered cheating if you obtain any kind of information about answers and solutions to any of the assignments in this course from any non-intended source (including your peers) or conversely transfer such information to others. When in doubt, ask the instructor. Nobody begins the semester with the intention of cheating. Students who cheat do so because they fall behind gradually, and then panic at the last minute. Some students get into this situation because they are afraid of an unpleasant conversation with an instructor if they admit to not understanding something. I would much rather deal with your misunderstanding early than deal with its consequences later. Please, feel free to ask for help as soon as you need it.  And remember: plagiarism violates academic honesty and Tufts faculty are required by Tufts policies to report any form of plagiarism.

Statement for Students with Disabilities:
 
The Americans with Disabilities Act (ADA) is a federal anti-discrimination statute that provides comprehensive civil rights protection for persons with disabilities. Among other things, this legislation requires that all students with disabilities be guaranteed a learning environment that provides for reasonable accommodation of their disabilities. If you believe you have a disability requiring an accommodation, please contact Disability Services.

This page is maintained by Matthias Scheutz.
Last revised on September 7, 2010.