Communication Quality in Anticipatory Vehicle Swarms: A Simulation-Based Model

Abstract

Recent research indicates the important role anticipation plays in the planning and deployment of autonomous multi-vehicle systems. The present study is devoted to building a simulation model of a swarm of autonomous ground-based vehicles. It is assumed that the vehicles perform collaborative surveillance and threat mitigation activities under difficult environmental conditions. Their performance is evaluated as the efficiency of threat mitigation during a single operation cycle, the total damage sustained by the vehicles during the operation, as well as a factor related to operational costs. We will take into account the vulnerability of the LAN communication under the different circumstances that may occur during the swarm operation. Based on the simulation model implemented in Matlab, it has been shown that vehicles endowed with anticipatory decision algorithms and organized in an anticipatory network perform considerably better compared to the behavior that follows a natural swarm benchmark algorithm. The advantage of an anticipatory network organization is particularly salient in case of communication disturbances. In summary, a smooth operation of the swarm can be ensured either by a reliable communication between vehicles via a local network or by implementing an anticipatory self-organization algorithm. Specifically, the latter can compensate for permanent communication deficiencies and may be particularly useful in case of its temporary fallouts due to unexpected disturbances.