Developing decentralized control for soft robots inspired by nature
Mannus Schomaker defended his PhD thesis at the Department of Mechanical Engineering on January 7th.
The robots in Mannus Schomaker鈥檚 research are inspired by animals like sea stars and worms. They coordinate their movements through physical interaction with each other and their surroundings, allowing them to move towards goals, adapt to damage, and respond to changing conditions without any central control. To recreate this in soft robots, Schomaker built a modular system in which each limb was a self-contained module that optimizes its behavior through a feedback loop based on limited sensing, short-term memory, and computation.
Coordinated movement
By harnessing the mechanical intelligence of soft pneumatic actuators, cyclic on鈥搊ff inputs to a pump at a fixed frequency were converted into complex bending and stepping motions. When multiple limbs were physically connected and each limb independently learned the phase of its oscillating motion, coordinated movement emerged.
Expanding and contracting to move
To gain a better understanding of how this coordination arises, Schomaker built a second modular system of self-contained units. In this system, the modules use the same strategy for sensing and processing, but their actuation is limited, making them immobile on their own. Instead, they expand and contract their connections to other physically linked modules on a two-dimensional plane. When interconnected in grid-like configurations, the system as a whole can break frictional symmetry to achieve locomotion, similar to earthworms that expand and contract segments.
Non-electronic movement
This research also shows that the control required to move soft robots does not need to be electronic. Rhythmic movement can be built directly into soft, air-driven components. Changing how these components are connected can quickly reshape how the robot moves. Simulations further showed that when body shape and control are designed together, even very simple robots can switch between different behaviors as their environment changes.
An alternative vision for the future
Overall, this research presents an alternative vision for robotics in which intelligence is not programmed from the top down but emerges from the body itself. By embedding functions into materials, using simple feedback, and leveraging interaction with the environment, soft robots can achieve robust, adaptable, and autonomous behavior.
This research was done at the research institute AMOLF.
Title of PhD thesis: . Supervisors: Dr. Bas Overvelde en Dr. Simon Eugster.