Designing collective behaviors for robot swarms is a difficult endeavor due to their fully distributed, highly redundant, and ever-changing nature. To overcome the challenge, a few approaches have been proposed, which can be classified as manual, semi-automatic, or automatic design. This paper is intended to be the manifesto of the automatic off-line design for robot swarms. We define the off-line design problem and illustrate it via a possible practical realization, highlight the core research questions, raise a number of issues regarding the existing literature that is relevant to the automatic off-line design, and provide guidelines that we deem necessary for a healthy development of the domain and for ensuring its relevance to potential real-world applications.

We investigate the automatic design of communication in swarm robotics through two studies. We first introduce Gianduja an automatic design method that generates collective behaviors for robot swarms in which individuals can locally exchange a message whose semantics is not a priori fixed. It is the automatic design process that, on a per-mission basis, defines the conditions under which the message is sent and the effect that it has on the receiving peers. Then, we extend Gianduja to Gianduja2 and Gianduja 3, which target robots that can exchange multiple distinct messages. Also in this case, the semantics of the messages is automatically defined on a per-mission basis by the design process. Gianduja and its variants are based on Chocolate, which does not provide any support for local communication. In the article, we compare Gianduja and its variants with a standard neuro-evolutionary approach. We consider a total of six different swarm robotics missions. We present results based on simulation and tests performed with 20 e-puck robots. Results show that, typically, Gianduja and its variants are able to associate a meaningful semantics to messages.

We introduce Gianduja, an automatic design method that generates communication-based behaviors for robot swarms. Gianduja extends Chocolate, a previously published design method. It does so by providing the robots with the capability to communicate using one message. The semantics of the message is not a priori fixed. It is the automatic design process that implicitly defines it, on a per-mission basis, by prescribing the conditions under which the message is sent by a robot and how the receiving peers react to it. We empirically study Gianduja on three missions and we compare it with the aforementioned Chocolate and with EvoCom, a rather standard evolutionary robotics method that generates communication-based behaviors. We evaluate the behaviors produced by the three automatic design methods on a swarm of 20 e-puck robots. The results show that Gianduja uses communication meaningfully and effectively in all the three missions considered. The aggregate results indicate that, on the three missions considered, Gianduja performs significantly better than the two other methods under analysis.

Neuro-evolution is an appealing approach to generating collective behaviors for robot swarms. In its typical application, known as off-line automatic design, the neural networks controlling the robots are optimized in simulation. It is understood that the so-called reality gap, the unavoidable differences between simulation and reality, typically causes neural network to be less effective on real robots than what is predicted by simulation. In this paper, we present an empirical study on the extent to which the reality gap impacts the most popular and advanced neuro-evolutionary methods for the off-line design of robot swarms. The results show that the neural networks produced by the methods under analysis performed well in simulation, but not in real-robot experiments. Further, the ranking that could be observed in simulation between the methods eventually disappeared. We find compelling evidence that real-robot experiments are needed to reliably assess the performance of neuro-evolutionary methods and that the robustness to the reality gap is the main issue to be addressed to advance the application of neuro-evolution to robot swarms.

Optimisation-based design is an effective and promising approach to realising collective behaviours for robot swarms. Unfortunately, the domain literature remains often vague on the exact role played by the human designer, if any. It is our contention that two cases should be disentangled: semi-automatic design, in which a human designer operates and steers an optimisation process (e.g., by fine-tuning the parameters of the optimisation algorithm); and (fully) automatic design, in which the optimisation process does not involve, need, or allow any human intervention. In the paper, we briefly review the relevant literature, we illustrate the hypotheses, the characteristics, and the core challenges of semi-automatic and automatic design, and we sketch the context in which they could be ideally applied.

We present Arlequin, an off-line automatic design method that produces control software for robot swarms by combining behavioral neural-network modules generated via neuro-evolution. The neuralnetwork modules are automatically generated once, in a mission-agnostic way, and are then automatically assembled into probabilistic finite-state machines to perform various missions. With Arlequin, our goal is to reduce the amount of human intervention that is required for the implementation or the operation of previously published modular design methods. Simultaneously, we assess whether neuro-evolution can be used in a modular design method to produce control software that crosses the reality gap satisfactorily. We present robot experiments in which we compare Arlequin with Chocolate, a state of the art modular design method, and EvoStick, a traditional neuro-evolutionary swarm robotics method. The preliminary results suggest that automatically combining neural-network modules into probabilistic finite-state machines is a promising approach to the automatic conception of control software for robot swarms.

Stigmergy is a form of indirect communication and coordination in which agents modify the environment to pass information to their peers. In nature, animals use stigmergy by, for example, releasing pheromone that conveys information to other members of their species. A few systems in swarm robotics research have replicated this process by introducing the concept of artificial pheromone. In this paper, we present Phormica, a system to conduct experiments in swarm robotics that enables a swarm of e-puck robots to release and detect artificial pheromone. Phormica emulates pheromone-based stigmergy thanks to the ability of robots to project UV light on the ground, which has been previously covered with a photochromic material. As a proof of concept, we test Phormica on three collective missions in which robots act collectively guided by the artificial pheromone they release and detect. Experimental results indicate that a robot swarm can effectively self-organize and act collectively by using stigmergic coordination based on the artificial pheromone provided by Phormica.

No abstract