Social learning in swarm robotics

In this paper, we present an implementation of social learning for swarm robotics. We consider social learning as a distributed online reinforcement learning method applied to a collective of robots where sensing, acting and coordination are performed on a local basis. While some issues are specific to artificial systems, such as the general objective of learning efficient (and ideally, optimal) behavioural strategies to fulfill a task defined by a supervisor, some other issues are shared with social learning in natural systems. We discuss some of these issues, paving the way towards cumulative cultural evolution in robot swarms, which could enable complex social organization necessary to achieve challenging robotic tasks. This article is part of a discussion meeting issue ‘The emergence of collective knowledge and cumulative culture in animals, humans and machines’.

Towards an integrated automatic design process for robot swarms

Background: The specification of missions to be accomplished by a robot swarm has been rarely discussed in the literature: designers do not follow any standardized processes or use any tool to precisely define a mission that must be accomplished. Methods: In this paper, we introduce a fully integrated design process that starts with the specification of a mission to be accomplished and terminates with the deployment of the robots in the target environment. We introduce Swarm Mission Language (SML), a textual language that allows swarm designers to specify missions. Using model-driven engineering techniques, we define a process that automatically transforms a mission specified in SML into a configuration setup for an optimization-based design method. Upon completion, the output of the optimization-based design method is an instance of control software that is eventually deployed on real robots. Results: We demonstrate the fully integrated process we propose on three different missions. Conclusions: We aim to show that in order to create reliable, maintainable and verifiable robot swarms, swarm designers need to follow standardised automatic design processes that will facilitate the design of control software in all stages of the development.

Remotely Controlled Robot Swarms: A Structural Analysis and Model for Structural Optimization

In this paper, the authors analyse the structure of robot swarms. Drive, communication, and artificial intelligence technologies have reached a level where the inspiration of the animal world has become very useful for the development of systems of which people have dreamed for centuries. A short introduction describes the need for and expectations of autonomous robots and vehicles. A classification of swarm structures is based on animals such as bees or ants. Two main types of swarms are recognized: structural (master–slaves) and non-structural. The operator controls both of them remotely. The swarm structure has a great influence on the structure of single robots. A computer model with an object programming definition was worked out, and a simulation of the presented swarm structure is provided. The results are described in the paper. The algorithm codes analysed in this paper are included in an appendix.

Almost global convergence to practical synchronization in the generalized Kuramoto model on networks over the n-sphere

From the flashing of fireflies to autonomous robot swarms, synchronization phenomena are ubiquitous in nature and technology. They are commonly described by the Kuramoto model that, in this paper, we generalise to networks over n-dimensional spheres. We show that, for almost all initial conditions, the sphere model converges to a set with small diameter if the model parameters satisfy a given bound. Moreover, for even n, a special case of the generalized model can achieve phase synchronization with nonidentical frequency parameters. These results contrast with the standard n = 1 Kuramoto model, which is multistable (i.e., has multiple equilibria), and converges to phase synchronization only if the frequency parameters are identical. Hence, this paper shows that the generalized network Kuramoto models for n ≥ 2 displays more coherent and predictable behavior than the standard n = 1 model, a desirable property both in flocks of animals and for robot control.

When less is more: Robot swarms adapt better to changes with constrained communication

To effectively perform collective monitoring of dynamic environments, a robot swarm needs to adapt to changes by processing the latest information and discarding outdated beliefs. We show that in a swarm composed of robots relying on local sensing, adaptation is better achieved if the robots have a shorter rather than longer communication range. This result is in contrast with the widespread belief that more communication links always improve the information exchange on a network. We tasked robots with reaching agreement on the best option currently available in their operating environment. We propose a variety of behaviors composed of reactive rules to process environmental and social information. Our study focuses on simple behaviors based on the voter model—a well-known minimal protocol to regulate social interactions—that can be implemented in minimalistic machines. Although different from each other, all behaviors confirm the general result: The ability of the swarm to adapt improves when robots have fewer communication links. The average number of links per robot reduces when the individual communication range or the robot density decreases. The analysis of the swarm dynamics via mean-field models suggests that our results generalize to other systems based on the voter model. Model predictions are confirmed by results of multiagent simulations and experiments with 50 Kilobot robots. Limiting the communication to a local neighborhood is a cheap decentralized solution to allow robot swarms to adapt to previously unknown information that is locally observed by a minority of the robots.

Secure and secret cooperation in robot swarms

The importance of swarm robotics systems in both academic research and real-world applications is steadily increasing. However, to reach widespread adoption, new models that ensure the secure cooperation of large groups of robots need to be developed. This work introduces a method to encapsulate cooperative robotic missions in an authenticated data structure known as a Merkle tree. With this method, operators can provide the “blueprint” of the swarm’s mission without disclosing its raw data. In other words, data verification can be separated from data itself. We propose a system where robots in a swarm, to cooperate toward mission completion, have to “prove” their integrity to their peers by exchanging cryptographic proofs. We show the implications of this approach for two different swarm robotics missions: foraging and maze formation. In both missions, swarm robots were able to cooperate and carry out sequential tasks without having explicit knowledge about the mission’s high-level objectives. The results presented in this work demonstrate the feasibility of using Merkle trees as a cooperation mechanism for swarm robotics systems in both simulation and real-robot experiments, which has implications for future decentralized robotics applications where security plays a crucial role.

Swarm robots in science fiction

In both science and science fiction, robot swarms are out of control.

Self-reconfigurable multilegged robot swarms collectively accomplish challenging terradynamic tasks

Swarms of ground-based robots are presently limited to relatively simple environments, which we attribute in part to the lack of locomotor capabilities needed to traverse complex terrain. To advance the field of terradynamically capable swarming robotics, inspired by the capabilities of multilegged organisms, we hypothesize that legged robots consisting of reversibly chainable modular units with appropriate passive perturbation management mechanisms can perform diverse tasks in variable terrain without complex control and sensing. Here, we report a reconfigurable swarm of identical low-cost quadruped robots (with directionally flexible legs and tail) that can be linked on demand and autonomously. When tasks become terradynamically challenging for individuals to perform alone, the individuals suffer performance degradation. A systematic study of performance of linked units leads to new discoveries of the emergent obstacle navigation capabilities of multilegged robots. We also demonstrate the swarm capabilities through multirobot object transport. In summary, we argue that improvement capabilities of terrestrial swarms of robots can be achieved via the judicious interaction of relatively simple units.