Contact & Learning:
Is the structure of contact still useful for solving
manipulation and locomotion challenges?
Contact & Learning:
Is the structure of contact still useful for solving
manipulation and locomotion challenges?
This workshop brings together researchers from different areas in robotics to explore the question of if the structure of contact is still valuable to learning-based control of robots. By contact, we mean the timing, location, force, and mode of physical interaction between the robot and an object or the environment. This structure has historically been critical to successful control of robot locomotion and manipulation tasks. However, modern learning-based solutions including reinforcement learning and imitation learning do not need to explicitly plan or reason about this structure. Thus, this workshop will address the key question: Is the structure of contact still useful for solving manipulation and locomotion challenges?
Should the robot plan for specific contact conditions or let them be implicitly defined by the control algorithm?
Can our knowledge of contact phenomenon be leveraged to build better learning-based controllers, e.g. by exploring the space of contact conditions?
Do contact conditions make for a good intermediate task representation to help with generalization to different tasks by explicitly modeling and preserving contacts?
Are differentiable simulations and gradients of contact useful in building more sample-efficient learning algorithms?
Submissions Open: June 25th, 2026, 13:00 CEST.
Submission Deadline: August 3rd, 2026, Anywhere on Earth.
Notification of Acceptance: August 10th, 2026.
Camera Ready Revision: September 1st, 2026.
Workshop Date: October 1st, 2026.
University of Michigan
Massachussetts Institute Technology
Carnegie Mellon University
Technical University of Darmstadt
University of Pennsylvania
Call for Abstracts:
We invite contributions from the community that address these questions. Submission details:
2 page maximum (excluding references)
IEEE/RSJ IROS double-column format (format available here)
All abstracts must explicitly address one of the workshop topics above! Please copy in one or more of the questions and provide your insight in the text of the abstract.
Anonymization: Initial submissions should be anonymized according to the IEEE RAS guidelines.
All accepted abstracts will be invited to present a poster at the workshop. The poster must also explicitly address one of the questions above somewhere on the poster.
Selected abstracts will also be invited to give an oral presentation at the workshop
All accepted abstracts will be made publicly available on the website. Note that this is not a full “publication”, it will not be assigned a DOI, etc. As such, we accept submissions of material in preparation, under review, or recently (within the last 6 months) published in other conferences or journals.
Submission Link: https://openreview.net/group?id=IEEE.org/IROS/2026/Workshop/Contact_Learning
Key Dates:
Submissions open: June 25th, 2026, 13:00 CEST.
Submission deadline: August 3rd, 2026, Anywhere on Earth.
Responses to authors: August 10th, 2026.
Camera ready version: September 1st, 2026.
Workshop date: October 1st, 2026.
We look forward to seeing your submissions and hearing your insights at the workshop in October!
1000 Fort Duquesne Boulevard
Pittsburgh, PA 15222
United States
Technical University of Munich
Technical University of Berlin
Technical University of Munich
Technical University of Munich
Technical University of Munich
Carnegie Mellon University