Disentangling automatic and semi-automatic approaches to the optimization-based design of control software for robot swarms

Abstract

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.

AutoMoDe-Mate: Automatic off-line design of spatially-organizing behaviors for robot swarms

“…The literature in these domains is extensive and therefore we restrict our attention to studies in swarm robotics [3,4]. In particular, we focus on studies that are relevant to the automatic design of robot swarms [8,40].…”

Reflections on the future of swarm robotics

“…First, the increasing complexity of swarm systems is such that their design cannot be accomplished solely by traditional approaches. The more robot swarms will be confronted with uncertain/unpredictable environments and will rely on intricate patterns of interactions, the more automated design methodologies will be necessary to obtain desired behaviors, as they can be employed to generate individual rules and evaluate them for their effects on swarm performance [9]. Machine learning with datadriven approaches becomes relevant whenever model-based solutions are too demanding, for instance when it is difficult to provide a precise model of the robot-environment interactions (e.g., due to complex physical interactions among micro/nano robots, or due to unpredictable environmental dynamics as caused by underwater currents).…”

Swarm SLAM: Challenges and Perspectives

“…Also, when working with robot swarms, one should consider how the control software of the individual robots will be designed. Studies have shown that the automatic off-line design of robot swarm can outperform manual design ( Birattari et al, 2019 ; Birattari et al, 2020 ) by building control software from simple atomic behaviors. A recent work in automatic design has also shown that exploration capabilities might come from the interaction between atomic behaviors and not only from the exploration schemes embedded in these atomic behaviors ( Spaey et al, 2020 ).…”