Towards Intelligent Active Particles

In this book chapter we describe recent applications of artificial intelligence and in particular machine learning to active matter systems. Active matter is composed of agents, or particles, that are capable of propelling themselves. While biological agents like bacteria, fish or birds naturally possess a certain degree of "intelligence", synthetic active particles like colloidal microswimmers and electronic robots can be equipped with different levels of artificial intelligence, either internally (as for robots) or via a dynamic external control system. This book chapter briefly discusses existing approaches to make synthetic particles increasingly "intelligent" and then focuses on the usage of machine learning to approach navigation and communication problems of active particles. Basic questions are how to steer a single active agent through a complex environment to reach or discover a target in an optimal way and how active particles need to cooperate to efficiently collect a distribution of targets (e.g. nutrients or toxins) from their complex environment.