Adapting to task changes without forgetting previous knowledge is a key skill for intelligent systems, and a crucial aspect of lifelong learning. Swarm controllers, however, are typically designed for specific tasks, lacking the ability to retain knowledge across changing tasks. Lifelong learning, on the other hand, focuses on individual agents with limited insights into the emergent abilities of a collective like a swarm. To address this gap, we introduce a lifelong evolutionary framework for swarms, where a population of swarm controllers is evolved in a dynamic environment that incrementally presents novel tasks. This requires evolution to find controllers that quickly adapt to new tasks while retaining knowledge of previous ones, as they may reappear in the future. We discover that the population inherently preserves information about previous tasks, and it can reuse it to foster adaptation and mitigate forgetting. In contrast, the top-performing individual for a given task catastrophically forgets previous tasks. To mitigate this phenomenon, we design a regularization process for the evolutionary algorithm, reducing forgetting in top-performing individuals. Evolving swarms in a lifelong fashion raises fundamental questions on the current state of deep lifelong learning and on the robustness of swarm controllers in dynamic environments.

Robot swarms have the potential to help groups of people with social tasks, given their ability to scale to large numbers of robots and users. Developing multi-human-swarm interaction is therefore crucial to support multiple people interacting with the swarm simultaneously - which is an area that is scarcely researched, unlike single-human, single-robot or single-human, multi-robot interaction. Moreover, most robots are still confined to laboratory settings. In this paper, we present our work with MOSAIX, a swarm of robot Tiles, that facilitated ideation at a science museum. 63 robots were used as a swarm of smart sticky notes, collecting input from the public and aggregating it based on themes, providing an evolving visualization tool that engaged visitors and fostered their participation. Our contribution lies in creating a large-scale (63 robots and 294 attendees) public event, with a completely decentralized swarm system in real-life settings. We also discuss learnings we obtained that might help future researchers create multi-human-swarm interaction with the public.

Robot swarms are decentralised systems that use simple rules to achieve collective goals, yet their real-world deployment is limited by a lack of understanding of human trust and perception. This study examines how swarm motion affects the trust of novice users in a service-oriented swarm, using an automated cloakroom as a test case. We conducted 20 human trials, where participants interacted with a swarm exhibiting either structured (grid-like) or organic (adaptive) motion, with performance controlled across conditions. Trust and perception were assessed via self-reporting questionnaires and eye-tracking data. Results indicate that performance and reliability, rather than motion, are the key drivers of trust. However, motion influenced perceived predictability, highlighting its role in designing transparent and user-friendly swarm systems.

MOSAIX is a new robot swarm platform built to be used in social settings. Consisting of up to 100 individual robot Tiles, MOSAIX is a social swarm system, aimed at helping humans in social tasks such as opinion-mixing and brainstorming. MOSAIX also has the potential to be used as a platform to study human-swarm interaction and swarm expressivity. MOSAIX is intended to be used outside laboratory settings and has already been used to collect 154 opinions about climate change in a local shopping mall, used by participants to create collaborative art and used as an educational tool for schoolchildren. We also discuss future applications, such as MOSAIX acting as smart post-it notes for ideation activities.

Decision-making among groups of humans can benefit from open discussion and inclusion of a diversity of opinions, promoting deliberative democracy. In this work, we test whether a swarm of robots can help facilitate decision-making by visually representing the diversity of opinions. We used a swarm of robots we built, called MOSAIX, that consists of 4-inch touchscreens-on-wheels robots called Tiles. The robots acted as physical avatars for opinions, helping them travel and mix together. We recruited 46 participants split into groups of 7 and 8 to test whether the robot movement had an impact on the decision-making process versus using the robots stationary in the participants' hands akin to smartphones. Furthermore, we wanted to test whether the participants felt comfortable expressing their opinion through the robots. Results show the participants indeed felt comfortable using the robots, and user engagement increased with the movement of the robots. The difference between the participants' first and last opinions also increased with the movement of the robots. We believe that robot swarms have not been used before to facilitate decision-making among a group of people. Therefore, our contribution is in testing the possibility of how and whether using a moving robot swarm helps humans reach a decision.

Due to the recent advances in robotics, large numbers of robots can be created that exhibit ‘swarm-like’ behavior. These robots, typically small and low-cost with restricted sensing, often exhibit Brownian motion similar to micro-particles. The development of algorithms that create collective behavior that is robust to external pressures has applications in outdoor exploration, search and rescue operations, and nanomedicine. Here, we outline how a swarm of minimal robots, exhibiting only Brownian motion and with limited sensing capabilities, can form trails using mechanisms inspired by diffusion-limited aggregation (DLA). We demonstrate how the trail is robust to obstacles and efficient at finding the closest target. We validate this algorithm both in simulation as well as in reality, using a swarm of up to 100 robots, and highlight the optimum requirements for trail formation.

In this paper, we investigate the best-of-n distributed decision problem in robot swarms. In this context, we compare the weighted voter model [25] with a three-valued model that incorporates an intermediate belief state meaning either `uncertain' or `indifferent'. We focus particularly on the trade-off between speed of convergence to a shared belief, and robustness to the presence of unreliable individuals in the population. By means of both simulation and embodied experiments in real robot swarms of 400 Kilobots, we show that the three-valued model is much more robust than the weighted voter model, but with decreased speed of convergence.

The success of swarm behaviors often depends on the range at which robots can communicate and the speed at which they change their behavior. Challenges arise when the communication range is too small with respect to the dynamics of the robot, preventing interactions from lasting long enough to achieve coherent swarming. To alleviate this dependency, most swarm experiments done in laboratory environments rely on communication hardware that is relatively long range and wheeled robotic platforms that have omnidirectional motion. Instead, we focus on deploying a swarm of small fixed-wing flying robots. Such platforms have limited payload, resulting in the use of short-range communication hardware. Furthermore, they are required to maintain forward motion to avoid stalling and typically adopt low turn rates because of physical or energy constraints. The tradeoff between communication range and flight dynamics is exhaustively studied in simulation in the scope of Reynolds flocking and demonstrated with up to 10 robots in outdoor experiments.

Aerial robots are often required to remain within the communication range of a base station on the ground to exchange commands, sensor data or as a safety mechanism. For this purpose, we propose a minimal control strategy for steering flying robots using only communication hardware (e.g. WiFi module or radio modem) instead of GPS or cameras. To avoid being dependent on the specifics of the communication hardware or its driver, we propose to measure the number of messages the robot receives from the base as a control input. Leashing is then performed by having the robot react to low message rates by moving towards the base in order to improve the communication. Results show both in theory and reality that this strategy can leash the robot to the base in scenarios with limited wind or base mobility.

It is generally challenging to design decentralized controllers for swarms of robots because there is often no obvious relation between the individual robot behaviors and the final behavior of the swarm. As a solution, we use artificial evolution to automatically discover neural controllers for swarming robots. Artificial evolution has the potential to find simple and efficient strategies which might otherwise have been overlooked by a human designer. However, evolved controllers are often unadapted when used in scenarios that differ even slightly from those encountered during the evolutionary process. By reverse-engineering evolved controllers we aim towards hand-designed controllers which capture the simplicity and efficiency of evolved neural controllers while being easy to optimize for a variety of scenarios.