Adaptive Application Behaviour for Robot Swarms using Mixed-Criticality

TL;DR

This paper proposes a mixed-criticality system model for swarm robotics that enables adaptive application behaviour in response to unreliable wireless communication, improving robustness in unpredictable network conditions.

Contribution

It introduces a novel integration of application and network layer behaviour using mixed-criticality systems, allowing robots to adapt dynamically to communication degradation.

Findings

Adaptive systems outperform static ones in unpredictable conditions

Mixed-criticality approach improves communication robustness

Simulation results confirm better support for unforeseen network issues

Abstract

Communication is a vital component for all swarm robotics applications, and even simple swarm robotics behaviours often break down when this communication is unreliable. Since wireless communications are inherently subject to interference and signal degradation, real-world swarm robotics applications will need to be able handle such scenarios. This paper argues for tighter integration of application level and network layer behaviour, so that the application can alter its behaviour in response to a degraded network. This is systematised through the implementation of a mixed-criticality system model. We compare a static application behaviour with that of an application that is able to alter its behaviour in response to the current criticality level of a mixed-criticality wireless protocol. Using simulation results we show that while a static approach is sufficient if the network conditions are known a priori, a mixed-criticality system model is able to adapt application behaviour to better support unseen or unpredictable conditions.