Automatic Off-Line Design of Robot Swarms: A Manifesto

Abstract

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.

Keywords: automatic design; collective behaviors; design methodology; evolutionary robotics; swarm robotics.

Figure 1

Fiorella's swarm gardening. Fiorella's robots can perform a large class of gardening missions. Through her website, customers book Fiorella's services, specify the interventions to be performed, and provide a description of their garden. On the basis of this information, while Fiorella drives her robots to customers, her algorithms automatically design and fine-tune the behavior of the robots so as to offer a tailored service. When she arrives at a customer location, the gardening swarm is operational and immediately deployed.

Figure 2

Flowchart diagram of the automatic off-line design process. A mission is sampled from a class of interest. Using computer-based simulations, an automatic design method defines a robot swarm tailored to the mission sampled. Once the automatic design terminates, the swarm is deployed in the target environment and has to cross the so-called reality gap—the possibly subtle but inevitable difference between simulation and reality (Brooks, ; Jakobi et al., 1995)—which is among the most challenging issues in automatic off-line design. The process is repeated ad libitum. It should be noted that an automatic design method could generate a robot swarm from scratch for every mission sampled or could refine and adapt a solution previously generated for a similar one. It could also, for example, produce a robot swarm by combining and modifying solutions (or partial solutions) contained in a catalog of template solutions that were pre-defined by a human expert. The only condition that needs to be respected for the process to be qualified as automatic off-line design is that such initial (partial) solutions must be selected without any per-mission human intervention and without recourse to tests performed in the target environment.