Robots that evolve on demand

Robert Baines

doi:10.1038/s41578-024-00711-z

Robots that evolve on demand

2024-09-12

0

PRE

AI

Rebecca Kramer‐Bottiglio *

DOI：10.1038/s41578-024-00711-z

Original

Original request for help

Save

Abstract

En 中文

Now more than ever, researchers are rethinking the way robots are designed and controlled - from the algorithms that govern their actions to the very atomic structure of the materials they are made from. In this Perspective, we collect and comment on recent efforts towards multipurpose machines that use shape-morphing materials and components to adapt to changing environments. To frame our discussion, we point out biological adaptation strategies that have been adopted by robots across different sizes and timescales. This contextualization segways into the notion of adaptive morphogenesis, which is formally defined as a design strategy in which adaptive robot morphology and behaviours are realized through unified structural and actuation systems. However, since its introduction, the term has been more colloquially used to describe 'evolution on demand'. We set out by giving examples of current systems that exhibit adaptive morphogenesis. Then, outlining projected key application areas of adaptive morphogenesis helps to scope the challenges and possibilities on the road to realizing future systems. We conclude by proposing performance metrics for benchmarking this emerging field. With this Perspective, we hope to spur dialogue among materials scientists, roboticists and biologists, and provide an objective lens through which we can analyse progress towards robots with rapidly mutable features that eclipse what is possible in biological processes.

Keywords:

LOOP DYNAMIC CONTROL

PHENOTYPIC PLASTICITY

SOFT

LOCOMOTION

EVOLUTION

MORPHOLOGY

ANIMALS

FUTURE

SENSOR

SKIN

Robots that evolve on demand

Abstract

Journal

Organization