Soft Robot
Soft robotics focuses on creating robots from flexible materials, enabling safer and more adaptable interaction with unstructured environments. Current research emphasizes developing accurate models for control, often employing neural networks (like recurrent neural networks and Echo State Networks), physical reservoir computing, and data-driven methods such as Lagrangian Operator Inference and Proper Orthogonal Decomposition for model reduction. This field is significant due to its potential applications in diverse areas like minimally invasive surgery, search and rescue, and underwater exploration, driving advancements in both robotics and materials science.
Papers
Learning Dynamic Tasks on a Large-scale Soft Robot in a Handful of Trials
Sicelukwanda Zwane (1), Daniel Cheney (2), Curtis C. Johnson (2), Yicheng Luo (1), Yasemin Bekiroglu (1 and 3), Marc D. Killpack (2), Marc Peter Deisenroth (1) ((1) UCL Centre for Artificial Intelligence, University College London, UK, (2) Department of Mechanical Engineering, Brigham Young University, USA, (3) Department of Electrical Engineering, Chalmers University of Technology, Sweden)
The untapped potential of electrically-driven phase transition actuators to power innovative soft robot designs
Diogo Fonseca, Pedro Neto
SoftSnap: Rapid Prototyping of Untethered Soft Robots Using Snap-Together Modules
Luyang Zhao, Yitao Jiang, Chun-Yi She, Muhao Chen, Devin Balkcom
Towards Reinforcement Learning Controllers for Soft Robots using Learned Environments
Uljad Berdica, Matthew Jackson, Niccolò Enrico Veronese, Jakob Foerster, Perla Maiolino