The GummiArm allows researchers in the embodied intelligence, cognitive and developmental robotics communities to implement models that could not be implemented in standard robotic arms.

A soft body like the GummiArm can outsource computation to the physical body, making the realisation of many tasks easier than when using classical control and computation models (Hauser, Füchslin & Pfeifer, 2014).

One of the motivations for developing the GummiArm was to run experiments on a system where intelligence emerges from the interaction between the body and the environment.

Another motivation was to put these concepts to work on real robotics applications where varying the passive compliance can have real benefits. For example for absorbing impacts by being soft when moving fast, and stiffening up when precision is required.

We hope you can join the GummiArm community!