ICRA 2024 Workshop on Co-design in Robotics: Theory, Practice, and Challenges

2024 International Conference on Robotics and Automation

Full-Day Workshop

Yokohama, Japan

Room G303 (North)

Thanks to everyone for a wonderful workshop!

We will be working on 2 major follow-ups to this workshop: a summary report of the discussion, and an RAL special issue on co-design.

If you would like us to keep you in the loop on either of these items, please let us know at: https://forms.gle/n8yiWUZPmDuVhXHDA

Overview

Robots are complex systems that must combine and integrate multiple coupled subsystems. The field of co-design seeks to understand how to jointly optimize robot subcomponents in order to produce a desired behavior. For example, one might consider how to jointly design the coupled kinematics and control architecture of the robot, taking into account their mutual dependencies, or how to design the robot mechanics and manufacturing methods. Co-design can also mean a process that engages end-users, stakeholders, and experts in the creation and development of robotic systems.

Robot co-design can offer various advantages – such as enhancing performance, efficiency, adaptability, usability, and user satisfaction – and there has been a large recent growth in both co-design methodologies and application areas. This workshop aims to gather researchers working in the various axes of co-design to discuss current trends, open problems, and challenges that the field is poised to tackle over the next 5-10 years.

Topics of interest include:

• Methodology

  • Theoretical foundations of co-design

  • Multi-objective co-design

  • Analytical approaches for co-design

  • AI/ML based co-design

  • Experimental verification of co-design approaches

  • Software frameworks for co-design

  • Interactive co-design

• Co-Design Scope

  • Manufacturing-aware co-design

  • Human-centered co-design 

  • Sensor / actuator co-design

  • Control / mechanism co-design

• Applications:

  • Legged locomotion

  • Soft robots

  • Canonical underactuated systems

  • Dexterous grippers

  • Flying robots

  • Underwater robots

Schedule

9:00 Welcome Remarks and Ground Rules
9:20 Short Talk: Audrey Sedal (McGill University)
"Co-Optimizing Design Control for Soft Robots in Contact with the Environment "
9:40 Short Talk: Matei Ciocarlie &  Zhanpeng He (Columbia University)
"(Co-)Designing Robots with Reinforcement Learning"
10:00 Coffee Break
10:20 Vignettes (10 min)
10:40 Tech Spotlights (2 min)
10:50 Breakout Discussion:
What are the technically challenging design areas that co-design should tackle over the next 5-10 years?
11:40 Reports and General Group Discussion
12:20 Lunch
14:00 Short Talk: Daniele  Pucci (Instituto Italiano di Tecnologia)
"Towards a Shared-Embodied Intelligence of Humanoid Robots: A Computational Human-Aware Co-Design Approach"
14:20 Short Talk: Guoxin Fang (Chinese University of Hong Kong)
"Improving Proprioception of Soft Robotics through Sensor/Actuator Co-Design: From Numerical Simulation to Data-Driven Methods"
14:40 Vignettes (10 min)
15:00 Tech Spotlights (2 min)
15:30 Coffee Break
16:00 Breakout Discussion:
What are the new areas of research / methods that will allow us to tackle these challenges?
17:00 Reports and General Group Discussion
17:30 Adjourn

Invited Speakers

Vignettes (Abstracts will be shared with participants at the workshop)

Morning Session (9:40 - 10:00)

• Gioele Zardini: Compositional Design of Autonomous Systems: From Hardware Selection to Decision Making

• Gabriele Fadini: Making Practical Use of Computational Optimization in Legged Robotics

Afternoon Session (14:40 - 15:00)

• Durgesh Haribhau Salunkhe, Guillaume Michel, Shivesh Kumar, Damien Chablat: Enhancing Otological Surgery: Co-Designing a Parallel Robot with Surgeon Input

• Wei-Hsi Chen, J. Diego Caporale, Daniel E. Koditschek, Cynthia R. Sung: Robogami Reveals the Utility of Slot-Hopper for Co-Design of DOQ's Body and Behavior 

Tech Spotlights (Abstracts will be shared with participants at the workshop)

Morning Session (10:30-10:40)

• Apoorv Vaish, Oliver Brock: Co-designing Manipulation Systems Using Task-Relevant Constraints

• Federico Girlanda, Lasse Shala, Shivesh Kumar, Frank Kirchner: Robust Co-Design of Canonical Underactuated Systems for Increased Certifiable Stability

• Mitchell B. Fogelson, Giusy Falcone, Zachary Manchester: CAPO: Control and Actuator Placement Optimization For Co-Design of High-DoF Nonlinear Robotic Systems

• Jolan Wauters, Tom Lefebvre, Guillaume Crevecoeur: Addressing the Reality Gap of Model-Based Aerial Robotics Development through Robust Co-Design

• Woohyun Cha, Daegyu Lim, Seonghee Kim, Jaeheung Park: The Design and Control of Bolt10: A Modification of The Open-Source Robot Bolt

Afternoon Session (15:00 - 15:30)

• Varun Madabushi, Katie M. Popek, Craig Knuth, Galen Mullins, Brian A. Bittner: Rapid Co-design of Task-Specialized Whegged Robots for Ad-Hoc Needs

• Gabriel Bravo-Palacios, Patrick M. Wensing: Robust Robot Co-Design: A Multi-scenario Perspective

• Jonathan Külz, Sehoon Ha, Matthias Althoff: Gradient-based Manipulator Morphology Optimization

• Juhyun Kim, Seungbin You, Eunho Sung, Dongjun Kim, Jaeheung Park: Analysis of Stair Climbing Bipedal Mechanism with 4-Bar linkage Knee Design

• Fabio Bergonti, Gabriele Nava, Valentin Wüest, Antonello Paolino, Giuseppe L'Erario, Daniele Pucci, Dario Floreano: Co-Design Optimisation of Morphing Topology and Control of Winged Drones

• Divij Grover, Adrian Buda, Sophie Armanini, Urban Fasel: Co-design Optimization using Deep Reinforcement Learning with Bayesian Optimization

• Yifang Zhang, Jingcheng Jiang, and Nikos G. Tsagarakis: On the Optimal Design of A Novel Passive Parallel Elastic Actuation Mechanism for Load Compensation in Legged Robots

• Lechen Zhang : CUDA-Accelerated Soft Robot Neural Evolution with Large Language Model Supervision

• Reed Truax, Feng Liu, Souma Chowdhury, Ryan St. Pierre: Optimising Design and Control of Running Robots Abstracted as a Torque Driven Spring Loaded Inverted Pendulum

• James M Bern: Conversation: An Extremely Fast and Simple CAD Program for 3D-Printed Parts and Robots

• Nitish Sontakke, K. Niranjan Kumar, Sehoon Ha: RoboDesignGPT: Design Optimization using LLMs

Workshop Organizers

This workshop is jointly organized by the TC on Mechanisms and Design and the TC on Model-Based Optimization for Robotics. 

TC on Mechanisms and Design

• Cynthia Sung, University of Pennsylvania

• Darwin Lau, Chinese University of Hong Kong

• Hannah Stuart, University of California, Berkeley

• Pauline Pounds, University of Queensland

TC on Model-Based Optimization for Robotics

• Shivesh Kumar, Chalmers University of Technology/DFKI

• Patrick Wensing, University of Notre Dame

Supported by: