Emergence of Consensus in a Multi-Robot Network: From Abstract Models to Empirical Validation

Abstract

Consensus dynamics in decentralised multiagent systems are subject to intense studies, and several different models have been proposed and analysed. Among these, the naming game stands out for its simplicity and applicability to a wide range of phenomena and applications, from semiotics to engineering. Despite the wide range of studies available, the implementation of theoretical models in real distributed systems is not always straightforward, as the physical platform imposes several constraints that may have a bearing on the consensus dynamics. In this paper, we investigate the effects of an implementation of the naming game for the kilobot robotic platform, in which we consider concurrent execution of games and physical interferences. Consensus dynamics are analysed in the light of the continuously evolving communication network created by the robots, highlighting how the different regimes crucially depend on the robot density and on their ability to spread widely in the experimental arena. We find that physical interferences reduce the benefits resulting from robot mobility in terms of consensus time, but also result in lower cognitive load for individual agents.

A gentle introduction to the minimal Naming Game

“…It has been studied in fully connected graphs (i.e. in mean-field or homogeneous mixing populations) [12,3,16], regular lattices [17,18], small world networks [19,18,20], random geometric graphs [21,18] and static [22,23,24,25], dynamic [26,27] and empirical [28,29] complex networks. It has been shown also that the final state of the system is always consensus [30], but stable polarized states can be reached introducing a simple confidence/trust parameter [31].…”

Embodied Evolution of Self-organised Aggregation by Cultural Propagation

“…It is possible to notice that the evolutionary model fails to produce stable aggregates when N = 25. This is because the MNG is particularly slow at low densities, because interactions among agents happen with very low probability [30,11]. As a consequence, the number of successful games is small-also due to mutations disturbing the language dynamics-and clusters quickly disband.…”

“…As a model of cultural evolution, we use the Naming Game (NG), which was developed to study the evolution of human language through statistical physics [2] and artificial life experiments [28]. The NG has actually already been studied within robotic swarms [30,11], whereby the swarm dynamics (e.g., random walk and aggregation) and the NG had a mutual effect on each other. However, to the best of our knowledge, there has been no attempt to use the NG as an embodied evolution approach.…”

Kilogrid: a novel experimental environment for the Kilobot robot

“…Charging the batteries of a large number of Kilobots can also be achieved simultaneously by moving all the Kilobots on a charging plate and sandwiching them with a top conductive grid, making it possible to apply a 6 VDC tension to the battery inputs (top connector and any of the Kilobot legs). Overall, the Kilobot design has solved several experimental issues that led to its adoption as a standard platform in several academic contexts, as demonstrated by successful experiments with tens or hundreds of Kilobots (e.g., collective decision making [28,30,32], collective transport [23] and manipulation [4] of objects, self-assembly [25,26] and disassembly [10]).…”