Structured interactions improve distributed coordination beyond model scaling in a real-world multi-robot system

2026-05-27

AI Summary

This study addresses the trade-off between interaction architecture and individual model scale in multi-robot systems under fixed hardware budgets. Conducting real-world transportation and mapping tasks with a fleet of ten physical robots, the authors systematically compare fully connected and modular hierarchical communication topologies while evaluating the impact of neural network size. For the first time in a physical multi-robot setting, empirical results demonstrate that adopting a modular hierarchical topology yields a 47-point improvement (on a normalized 0–100 scale) in system performance, whereas doubling the hidden layer size confers at most a 9-point gain. Mixed-effects modeling further confirms the dominant influence of communication topology over model scale. These findings underscore the critical role of system-level interaction design in enhancing multi-agent coordination performance.

Abstract

Scaling individual robot capabilities is common but costly. Here we investigate a system-level design question in real-world multi-robot coordination: given matched hardware budgets, does restructuring communication among robots yield larger gains than increasing onboard model size? Using a representative transport-and-mapping task with 10 physical robots (5 runs per condition, 60 runs total), we find that switching from fully connected to modular hierarchical interactions improves normalised performance by 47 points (0--100), whereas doubling neural network hidden size yields at most 9 points. Nested mixed-effects model comparisons show a substantially larger improvement in model fit for topology than for scale. The pattern is confirmed in independent SMAC replications; heterogeneous benchmark reanalyses provide secondary supporting consistency checks rather than primary evidence. Performance saturation beyond 1024 hidden units is observed in simulation-calibrated extrapolation, not directly on hardware. These results indicate that interaction structure can play a dominant role within the tested system and task setting, while broader quantitative generalisation remains to be established.

Problem

• multi-robot coordination
• communication topology
• model scaling
• distributed coordination
• system-level design

Innovation

• structured interaction
• multi-robot coordination
• communication topology
• model scaling
• modular hierarchy