Directed light beams are a promising means of control for microscopic agents, whether they are microrobots or phototatic microorganisms such as Volvox and ciliates. Given the simple reactive behaviors common to most microagents, there is likely to be a certain universality in light-beam algorithms that can usefully ‘herd’ such collectives around. Here, we develop three light-beam control algorithms to herd light-sensitive agents around a two-dimensional environment, each making varying assumptions about agent behavioral capacities. We test them with small swarms of Kilobot robots, which are about 3cm in size. These robots are convenient macro-scale demonstrators of possibilities at the micro-scale. The algorithms are tested in simulation and found to achieve the desired herding goals. Waypoint following missions were implemented using single robots and multiple robots to demonstrate more complex trajectories and highlight the effects of multiple robots interacting. One of the algorithms was tested with real robots and is shown to perform well, owing to good robustness to projection inaccuracies. Future swarm engineers could refer to a common toolbox of broadly effective light-based swarm control algorithms, which can be selected according to agent capabilities.

The effective control of microscopic collectives has many promising applications, from environmental remediation to targeted drug delivery. A key challenge is understanding how to control these agents given their limited programmability, and in many cases heterogeneous dynamics. The ability to learn control strategies in real time could allow for the application of robotics solutions to drive the behaviour of microscopic collectives towards desired outcomes. Here, we demonstrate Q-learning on the closed-loop Dynamic Optical Micro-Environment (DOME) platform to control the motion of light-responsive Volvox agents. The results show that Q-learning is efficient in autonomously learning how to reduce the speed of agents on an individual basis.

Robotic Canvas combines swarm robotics, human-robot interaction, and art. The project creates a robotic canvas using a swarm of decentralised robots that a human can interact with to paint. Beauty emerges from the robots interacting with one another, their environment, and a human performer. Robotic Canvas is a dynamic and interactive visual art medium, capable of displaying static images or video feeds, upon which humans can paint with their physical gestures or dedicated robots. Art making therefore becomes a collaborative and immersive performance between artists, digital content, and robotic agents. Results are demonstrated through a 200-robot performance, with each robot acting as a single pixel of the canvas. We further describe and characterise the various interaction modes, called “Painting Modes”, of the system. Results from this work form the basis for future research in expressive human-swarm interactions.

Robotic swarms rely on local communication between agents to exhibit cooperative emergent behaviours. Local communication is typically implemented with technologies that require dedicated electronics, that can be expensive and difficult to miniaturise or mass-produce. Computational resources are then needed to transform this information into a robot action following a set of rules, further limiting swarm lifetime (battery) and scalability. In this paper, we propose an alternative approach by using the concept of morphological computation (computation through morphology) for local communication in swarms. In such a swarm, local communication is implemented as simple mass-spring-damper systems between agents, instead of electronics. We test this approach in a simple scenario where a swarm has to squeeze through a narrow gap while floating on water. We tested different types of swarms (with different levels of control) and measured their average performance and energy efficiency. We found that by offloading the majority of communication and information processing to the morphology, swarms can exhibit interesting, emergent, cooperative behaviour to solve the given task.

Robot swarms can solve tasks that are impossible or too hazardous for single robots. For example, following a nuclear radiation leak, a user may wish to establish a distributed communication chain that partly extends into the most dangerous areas to gather new information. The challenge is to create long chains while maintaining chain connectivity (‘connected reach’), where those at the distant end of the chain are more likely to be disconnected. Here we take the concept of dynamic ‘boldness’ levels from animal behavior ( Stegodyphus social spiders) to explore such risky environments in a way that adapts to the size of the group. Boldness is implemented as a continuous variable associated with the risk appetite of individuals to explore regions more distant from a central base. We present a decentralized mechanism for robots, based on the frequency of their social interactions, to adaptively take on ‘bold’ and ‘shy’ behaviors. Using this new bioinspired algorithm, which we call SPIDER, swarms are shown to adapt rapidly to the loss of bold individuals by regenerating a suitable shy–bold distribution, with fewer bolder individuals in smaller groups. This allows them to dynamically trade-off the benefits and costs of long chains (information retrieval versus loss of robots) and demonstrates the particular advantage of this approach in hazardous or adversarial environments.

Infiltrating a swarm of artificial or living agents using a single monitoring robot could allow for the assessment of their swarm rules and parameters without the need for any external infrastructure. The inferred swarm model could then be used to control these swarms, for example to guide them to safe areas. In this study we introduce a scheme for autonomous artificial agents to extract knowledge about the interactions within a swarm of interest. By infiltrating the swarm of interest with a monitoring robot and constantly measuring the distance between the infiltrator and its nearest neighbour, the repulsion radius of the swarm agents can be estimated. Though this method works for a range of tested parameters, it is still limited to a specific model of interaction.

Swarm engineering allows us to design self-organised systems across scales, from trillions of nanoparticles for cancer treatment, to thousands of robots for environmental monitoring. Scaling to such large numbers requires discovering new collective behaviours that rely largely on random motion and simple communication between agents and their environment.