Learning to Design 3D Printable Adaptations on Everyday Objects for Robot Manipulation

Abstract

Advancements in robot learning for object manipulation have shown promising results, yet certain everyday objects remain challenging for robots to effectively interact with. This discrepancy arises from the fact that human-designed objects are optimized for human use rather than robot manipulation. To address this gap, we propose a framework to automatically design 3D printable adaptations that can be attached to hard-to-use objects, thus improving "robot ergonomics". Our learning-based framework formulates the adaptation design and control as a dual Markov decision process and is able to improve robot-object interactions for various robot end effectors and objects. We further validate our designs in the real world with a Franka Panda robot.

Framework

Overview of our framework. We formulate the design and control as a dual Markov decision process and adopt a reinforcement learning framework to jointly optimize them.

Comparison of Design Strategies

Learning curves for varying design strategies. Random designs outperform trying to use the original object with no design, but the learned designs are the best. Shaded regions indicate standard error across a minimum of three random seeds.

Supplementary Video

BibTeX

@article{guolearning,
      author        = {Guo, Michelle and Liu, Ziang and Tian, Stephen and Xie, Zhaoming and Wu, Jiajun and Liu, C Karen},
      title         = {Learning to Design 3D Printable Adaptations on Everyday Objects for Robot Manipulation},
      booktitle     = {IEEE International Conference on Robotics and Automation (ICRA)},
      year          = {2024},
      organization  = {IEEE}
    }