EECS227: Robot Perception

Overview:

Robot Perception is the cornerstone of modern robotics, enabling machines to interpret, understand, and respond to an array of sensory information they encounter.  In this course, students will learn the basic principles of typical sensor hardware on a robotics system (e.g., vision, tactile, and acoustic sensors), the algorithms that process the raw sensory data, and how to make actionable decisions from that information. We will discuss both modular-based perception systems and end-to-end policy learning.

Throughout the course, students will incrementally build their own vision-based robotics system in simulation via a series of homework coding assignments. Students will also get hands-on experience in the iterative improvement process for robotics system design, which includes testing and evaluating the developed system, identifying issues and weaknesses of the system, and making improvements to the system based on evaluations.

Instructor: Shuran Song

TA:  Chuer Pan, Zhanyi Sun, Ben Pekarek

Time:  10:30a - 11:50a MW

• Location:   SKILLAUD

• Credit: 3 or 4 credits 

• Office Hours:  

TA office hours. Questions  related to homework:
In-person: Mon 4:00 - 6:00 PM.  Location: Gates 100. 

 Online Zoom: Tues 6:00 - 8:00 PM. Zoom link, also found on Canvas under the Zoom tab. 

Note: Week of Nov 3, Tuesday is university holiday, so online OH is rescheduled to Thursday, 6-8 pm. Zoom link.

Shuran's office hours: Questions  related to lecture:
Wed 12:00 - 1:00 pm (After lecture outside SkillAUD)

• Lecture Slides: https://www.dropbox.com/scl/fo/wc215mghdz5bgr52i1u3z/ABoj8dtt7C3wMsMV4rYsaDg?rlkey=u09rnhhjrpheqwv2k0ql8oiac&dl=0 (Dropbox)

• Lecture Recordings: on Canvas

• Ed Discussion on Canvas

Pre-requisites:

• Data Structures

• Knowledge of  Python. We will be using Python extensively in this course.

• Knowledge of elementary linear algebra (e.g. MATH 51)

We will use matrix transpose, inverse, and other operations to do algebra with matrix expressions. We’ll use transformation matrices to rotate/transform points. These topics are important for the homework.  

• If you are not sure whether you are ready for the course, please check with the course instructor. 

Grading:

• 5 HW (15 % each)  + 1 Quizes (25%)

• 4-credit students are required to submit an additional 2-page report in the last assignment: 

Write a literature review or survey on topics we discussed in lecture. For example, a survey on "single-view depth estimation." In the survey, please categorize the works into three categories.  For each category, describe the general idea, the work that belongs to this category, and the general pros and cons of this category of approach. 

• For students joining the course after HW1 deadline, we waive this homework score and add the weight of this homework to your final exam weighting, so the final will be 40% of your course grade rather than 25%.

Late Policy: 

• DDL is always at 5 pm. 

• You have 2 late days for the whole quarter (count by day, e.g., an hour late == a day)

• No exceptions beyond the 2 late days.  

• No more TA support after the deadline. If you choose to use late days, there will be no TA support for those days. 

Not seeing anything above? Reauthenticate