fer to specify the theme of the project, such as to build
a service robot or to address a problem with a tangible impact to society. The theme becomes the seed
for their ideas, driving their work.
With these open-ended requirements, each team
needs frequent feedback on their projects every one
to two weeks, both from the instructor as well as
from the other teams. In the robotics course, students
responded positively to having deadlines and
demonstrations every two weeks to drive their
progress. Weekly in-class status reports from all
groups and the use of an agile development methodology both encourage progress and invite feedback,
as well as allow problems and issues to be addressed
early. Often, students become bogged down in minor
issues that can consume extraordinary amounts of
time. Warn students to watch out for this, and then
use these weekly check-ins to detect such problems.
It is also extremely important to have a flexible syllabus. As the projects develop, you will likely need to
add or change topics to address specific needs of the
projects. My preference is to be upfront with the students from the first class that the course syllabus and
schedule will be highly dynamic. They will then
expect and accept changes easily, instead of protesting when the schedule is adjusted. Students in project-driven courses should have influence over the
course’s direction, so request feedback frequently
from students on the course schedule. Reading summaries help ensure that students are keeping up to
date with the schedule; having them due electronically with a hard deadline in the evening before each
class will help guarantee that students are prepared
to participate in the seminar discussion.
Project-driven courses are an amazing experience.
Consider teaching integrated AI through one of these
courses, and enjoy the many benefits for both you
and your students.
1. See A Roadmap for US Robotics: From Internet to Robotics 2013 Edition, Robotics Virtual Organization. (robotics-
3. See the Defense Advanced Research Projects Agency’s
page on the Heilmeier Catechism, www.darpa.mil/work-with-us/heilmeier-catechism
Aronson, E.; Blaney, N.; Stephan, C.; Sikes, J.; and Snapp, M.
1978. The Jigsaw Classroom. Beverly Hills, CA: Sage Publish-
Brachman, R. J. 2006. (AA)AI More Than the Sum of Its
Parts. AI Magazine
27( 4): 19–34.
Eaton, E.; Mucchiani, C.; Mohan, M.; Isele, D.; Luna, J. M.;
and Clingerman, C. 2016. Design of a Low-Cost Platform
for Autonomous Mobile Service Robots. Paper presented at
the IJCAI- 16 Workshop on Autonomous Mobile Service
Robots, New York, New York, 11 July.
Gomes, C. 2009. Computational Sustainability: Computa-
tional Methods for a Sustainable Environment, Economy,
and Society. National Academy of Engineering. The Bridge
on Frontiers of Engineering
Murphy, R. R. 2015. Meta-Analysis of Autonomy at the
DARPA Robotics Challenge Trials. Journal of Field Robotics
32( 2): 189–191. doi.org/10.1002/rob.21578
SRI International. 1969. SHAKEY: Experimentation in Robot
Learning and Planning (Video). Menlo Park, CA: SRI Interna-
van Beek, L.; Chen, K.; Holz, D.; Matamoros, M.; Rascon, C.;
Rudinac, M.; Ruiz des Solar, J.; and Wachsmuth, S. 2015.
RoboCup@Home 2015: Rules and Regulations (online). Palo
Alto, CA: Robocup Federation. ( www.robocupathome.org/
Wollowski, M.; Selkowitz, R.; Brown, L. E.; Goel, A.; Luger,
G.; Marshall, J.; Neel, A.; Neller, T.; and Norvig, P. 2016. A
Survey of Current Practice and Teaching of AI. The Sixth
Symposium on Educational Advances in Artificial Intelli-
gence (EAAI- 16). In Proceedings of the Thirtieth AAAI Confer-
ence on Artifcial Intelligence. Palo Alto, CA: AAAI Press.
Eric Eaton is a faculty member in the Department of Computer and Information Science at the University of Pennsylvania, and a member of the General Robotics, Automation, Sensing, and Perception (GRASP) lab. Prior to joining
Penn, he was a visiting assistant professor in the computer
science department at Bryn Mawr College, and a senior
research scientist at Lockheed Martin Advanced Technology Laboratories. His primary research interests lie in the
fields of machine learning, AI, and data mining with applications to service robotics, environmental sustainability,
and medicine. In particular, his research focuses on developing lifelong machine-learning systems that learn numerous tasks over a lifetime of experience in complex dynamic
environments, transfer learned knowledge to rapidly
acquire new abilities, and collaborate effectively with
humans and other agents.