CISC849 F2023

From class_wiki
Revision as of 13:51, 31 October 2023 by Cer (talk | contribs) (Schedule)
Jump to: navigation, search

Course information

Title CISC849-015 Robot Navigation and Autonomy
Description A hands-on approach to implementing mobile robot algorithms on a small wheeled platform, both in simulation and reality. We will focus on image- and depth-based sensing algorithms for obstacle segmentation, object recognition, and motion planning, including deep learning techniques in these areas.
When Tuesdays and Thursdays, 2:20-3:40 pm. When there is a homework due, no more than the first 30 minutes of each class will be in lecture format. The rest of the class period (and optionally the subsequent office hours) will be spent working on the robots.
Where Smith 211
Instructor Christopher Rasmussen, 446 Smith Hall, cer@cis.udel.edu
Office hours Tuesdays and Thursdays, 3:40-5 pm in Smith 211 (starting Aug. 31)
Grading
  • 60% 5 programming homeworks (12% each)
  • 15% 3 quizzes
  • 25% Final project

Programming assignments will be graded on how many of the subtasks you complete or demonstrate successfully.

For the overall course grade, a preliminary absolute mark will be assigned to each student based on the percentage of the total possible points they earn according to the standard formula: A = 90-100, B = 80-90, C = 70-80, etc., with +'s and -'s given for the upper and lower third of each range, respectively. Based on the distribution of preliminary grades for all students (i.e., "the curve"), the instructor may increase these grades monotonically to calculate final grades. This means that your final grade can't be lower than your preliminary grade, and your final grade won't be higher than that of anyone who had a higher preliminary grade.

I will try to keep you informed about your standing throughout the semester. If you have any questions about grading or expectations at any time, please feel free to ask me.

Academic policies Programming projects should be demo'd in class on the deadline day and uploaded to Canvas by midnight of that day (with a grace period of a few hours afterward...after sunrise is definitely late). A late homework is a 0 without a valid prior excuse. To give you a little flexibility, you have 3 "late days" to use on homeworks to extend the deadline to the next class period without penalty. No more than one late day may be used per assignment. Late days will automatically be subtracted, but as a courtesy please notify the instructor in an e-mail of your intention to use late days before the deadline.

Assignment submissions should consist of a directory containing all code (your .cpp/.py files, etc.), any output data generated (e.g., images, movies, etc.), and an explanation of your approach, what worked and didn't work, etc. contained in a separate text or HTML file. Do not submit executables or .o files, please! The directory you submit for each assignment should be packaged by tar'ing and gzip'ing it or just zip'ing it.

Students can discuss problems with one another in general terms, but must work independently or within their teams as specified for each assignment. This also applies to online and printed resources: you may consult them as references (as long as you cite them), but the words and source code you turn in must be yours alone. The University's policies on academic dishonesty are set forth in the student code of conduct here.

Instructions/Resources

Robot

Yes, our robot platform is a Roomba. Except it can't clean.

  • Platform: iRobot Create 3 running ROS2 Humble
  • SBC: Raspberry Pi 4 B running Ubuntu 22.04 and ROS2 Humble
  • Lidar: Slamtec RPLIDAR A1
  • Camera: RealSense D435
ROS

We are using ROS2, to be exact, and the Humble Hawksbill version. My laptop and the Raspberry Pi's on the robots are running Ubuntu Desktop 22.04 LTS, and I strongly recommend that you do the same. I say this not because of convenience, but after hard experience last semester with another group of students. Trying to install and run ROS2 on Windows or MacOS natively is a recipe for frustration and disappointment. Here are installation instructions for Ubuntu. Several students with MacOS had success using Ubuntu in a VM such as UTM. With a VM there are more issues such as setting up bridging mode for wifi, but it's workable.

A lot of public sample code and tutorials are available in both C++ and Python, but expect mostly C++ examples from me. I am agnostic about which of these two languages you use for homeworks and projects, but you will get your best support from me in C++.

Readings
  • K. Lynch and F. Park, Modern Robotics (MR), 2019
  • P. Corke, Robotics, Vision, and Control (RVC), 2017
  • K. Astrom and R. Murray, Feedback Systems (FS), 2008
  • S. Lavalle, Planning Algorithms (PA), 2006
Gazebo

This is a 3-D robot simulator


Schedule

Note: The blue squares in the "#" column below indicate Tuesdays.

# Date Topic Notes Readings/links Assignments/slides
1 Aug. 29 Introduction Background, course information slides
2 Aug. 31 hello robot, hello ROS ROS basics, Create ROS interface Creating & building ROS2 packages, ROS2 nodes,
ROS2 topics, Create 3 topics
slides
HW #1
3 Sep. 5 Kinematics/Dynamics Degrees of freedom, configuration space; wheeled systems (unicycle/car vs. differential drive) MR Chap. 2, PA Chap. 13.1.2 slides
4 Sep. 7 ROS ROS 2 workspaces and packages, subscribing and publishing, and timer callbacks slides
HW #1 due
5 Sep. 12 HW #2 HW #2
6 Sep. 14 Controllers Waypoint following, line following, trajectory and wall following (pure pursuit) RVC Chap. 4-4.1.2 slides
7 Sep. 19 HW #2
8 Sep. 21 Robot geometry URDFs, ROS rviz2 and tf2, odometry rviz2 user manual (Turtlebot3), Introduction to tf2, MR Chap. 13.4 slides
HW #2 due
9 Sep. 26 Costmaps and discrete motion planning Connecting to the lidar sensor; representing the environment as a map and basic planning rplidar_ros; MR Chap. 10.4,
Abbeel slides (skip SLAM, reflection maps)
slides
Quiz #1
HW #3
Isaac sim udbot Sep. 22 version
10 Sep. 28 Estimation Least-squares line-fitting and
outlier rejection with RANSAC for lidar scans
Scikit-learn ordinary least-squares and RANSAC slides
11 Oct. 3 HW #3
12 Oct. 5 Localization Particle filters, MCL Thrun particle filtering slides slides
13 Oct. 10 HW #3 HW #3 due
14 Oct. 12 Computer vision Connecting to the RGB-D camera, line finding realsense-ros Github, basic OpenCV color processing, Hough lines slides
HW #4
15 Oct. 17 Computer vision Tags/fiducials AprilTag Github, apriltag_ros slides
16 Oct. 19 HW #4 Coding time
17 Oct. 24 Computer vision Getting depth/3-D point cloud from the RealSense;
object detection
realsense-ros Github, perception_pcl for further analysis
YOLOv5 (Python, lots of requirements), Training YOLOv5 on custom data
slides
18 Oct. 26 HW #4 Coding time Quiz #2
HW #4 due
19 Oct. 31 HW #5 HW #5
20 Nov. 2 SLAM ROS slam_toolbox package slam_toolbox Github, Thrun FastSLAM slides
21 Nov. 7 HW #5 HW #5 due
22 Nov. 9 Motion learning Dynamical tasks, reinforcement learning Gymnasium Github
23 Nov. 14 Motion learning Final project
Nov. 16 NO CLASS
Instructor away
Nov. 21 NO CLASS
Thanksgiving break
Nov. 23 NO CLASS
Thanksgiving away
24 Nov. 28 Ethical & societal issues
25 Nov. 30 Ethical & societal issues Quiz then coding time Quiz #3
26 Dec. 5 Final project Coding time
27 Dec. 7 Final project demos/presentations Final project due