COMP 655: Advanced Topics in Robotic Manipulation

Instructor: Kaiyu Hang
Credits: 1 or 3
Location: DH 1042
Lectures: 4:00PM - 5:30PM W(23-AUG-2021 - 3-DEC-2021)
Office Hours: By appointment

Course Description

Manipulation is one of the major research fields in robotics. The ultimate goal is to enable robots, using any parts of their embodiments, to physically interact with and reconfigure the world. Being an interdisciplinary problem, research in robotic manipulation has to address a set of subproblems from various perspectives, such as grasping, motion planning, prehensile and non-prehensile manipulation, affordance and task modeling, hand design, and perception. This course not only provides a comprehensive review of state-of-the-art research on developing robotic manipulation systems of different functionalities and complexities, but also offers opportunities for students to formulate their own research problems, which are based on optional projects that can potentially lead to research publications. Due to its interdisciplinary nature, this course is designed and offered to students in Computer Science, Electrical and Computer Engineering, and Mechanical Engineering.

Prerequisites

No specific previous courses is required for taking this course. However, it will be helpful if you have some background knowledge in any of: AI, planning, machine learning, computer vision, control, optimization, sensing, kinematics, and dynamics.

Grading

1-credit: The final grade will be based on the weekly paper reviews (40%), in-class discussion and presentations (40%), and the final presentation (20%).
3-credit: The final grade will be based on the weekly paper reviews (20%), in-class discussion and presentations (20%), final presentation (15%), and the optional project (45%).


Writing Reviews

Two papers (research articles or survey/review papers) will be given one week before each class (except for the first week). Every student needs to read both papers and write a review for one of them. The review submission is due at 6pm on the day before each class. For writing the review, please follow the structure below:

Research Articles:
1. What problems are addressed in the paper?
2. What related work were discussed in this paper?
3. What is the core methodology of this work?
4. What is the major weakness of this work?
5. Do you think the experiments were well designed and the evaluations were informative?
6. What are potential future directions of this work?

Survey/Review Papers:
1. What is the research area being surveyed? Why is this area important?
2. What are the major problems in this field?
3. What are the major research methodologies in this field?
4. Which research methodology do you like the most? Provide your reasons.
5. Which research methodology you do not like? Provide your reasons.
6. What potential real-world applications can be developed based on the research of this field?

Presentations

In-class presentation: There will be two presentations (15 minutes) in every class on the two papers we read in the previous week. After each presentation, we will form two groups, one support and one against the paper, to hold a 30 minutes discussion. The presenter is responsible for initiating the debate.

Final Presentation: Every student will give a final presentation (10 minutes) to propose a research project, which has to address a relevant research question using novel ideas that can potentially make a contribution to the robotic manipulation community.

Optional Project (3-credit)

The optional project is research-oriented. In the first 3 weeks, students who would like to do such projects need to propose research ideas and discuss with me. Once we agree on a project topic, the student will conduct a complete research cycle, including reading relevant papers, formulate research questions, propose methods, design and conduct experiments, and finally write a 6-page report (template will be provided). The final presentation for 3-credit students will be 20 minutes.

Students can do optional projects individually or in groups of maximum 2 members.

Rice Honor Code

In this course, all students will be held to the standards of the Rice Honor Code, a code that you pledged to honor when you matriculated at this institution. If you are unfamiliar with the details of this code and how it is administered, you should consult the Honor System Handbook at http://honor.rice.edu/honor-system-handbook/. This handbook outlines the University's expectations for the integrity of your academic work, the procedures for resolving alleged violations of those expectations, and the rights and responsibilities of students and faculty members throughout the process.

Disability Resource Center

If you have a documented disability or other condition that may affect academic performance you should: 1) make sure this documentation is on file with the Disability Resource Center (Allen Center, Room 111 / adarice@rice.edu / x5841) to determine the accommodations you need; and 2) talk with me to discuss your accommodation needs.

Absence Policies

Please read Rice's Attendance and Excused Absences

Course Schedule (Tentative)

Sign up for your in-classl presentations via this doodle form
Week Topic Paper #1 Paper #2 Note
August 25, 2021 Introduction Trends and Challenges in Robot Manipulation Optional project start
September 1, 2021 Robot Grasping Minimum volume bounding box decomposition for shape approximation in robot grasping Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics
September 8, 2021 Hand Design, Control, and Synergies A Century of Robotic Hands Dimensionality reduction for hand-independent dexterous robotic grasping
September 15, 2021 Motion Planning Search-based planning for manipulation with motion primitives Probabilistic roadmaps for path planning in high-dimensional configuration spaces Optional project proposed and proposal submitted
September 22, 2021 Hand-Object Systems An overview of dexterous manipulation Haptic object recognition for multi-fingered robot hands
September 29, 2021 In-Hand Manipulation Manipulation for self-Identification, and self-Identification for better manipulation Learning dexterous in-hand manipulation
October 6, 2021 Nonprehensile Manipulation Nonprehensile Dynamic Manipulation: A Survey Nonprehensile whole arm rearrangement planning on physics manifolds
October 13, 2021 Sensing in Manipulation More Than a Feeling: Learning to Grasp and Regrasp Using Vision and Touch Enhancing visual perception of shape through tactile glances
October 20, 2021 Interactive Perception Interactive Perception: Leveraging Action in Perception and Perception in Action Interactive segmentation, tracking, and kinematic modeling of unknown 3D articulated objects
October 27, 2021 Multi-modal Manipulation Informing Multi-Modal Planning with Synergistic Discrete Leads Multi-modal motion planning in non-expansive spaces 1. Start writing proposal for final presentation
2. Optional project progress presentations
November 3, 2021 Representations A finite state machine based on topology coordinates for wrestling games A Framework for Optimal Grasp Contact Planning Final presentation topics proposed and proposal submitted
November 10, 2021 Dual-arm Manipulation Dual arm manipulation - A survey Humanoid motion planning for dual-arm manipulation and re-grasping tasks
November 17, 2021 Mobile Manipulation An Optimization Approach to Planning for Mobile Manipulation A constraint-based method for solving sequential manipulation planning problems Final presentations
December 1, 2021 Optional project Optional project presentations (Final report due December 3)