Home /Research /Virtual pheromones to control mobile robots. A neural network approach
SWARM

Virtual pheromones to control mobile robots. A neural network approach

Ioan Șușnea, Grigore Vasiliu, Adrian Filipescu, Adriana Şerbencu, A. Radaschin

Year
2009
Citations
5

Abstract

This paper introduces the concept of ldquovirtual pheromonesrdquo for use in controlling autonomous mobile robots. Virtual pheromones are stored in a map of the environment maintained by a ldquopheromone serverrdquo, which acts like a shared memory for all the agents. Autonomous agents communicate with the pheromone server by means of a radio communication link. A model of the virtual pheromone, which includes spatial diffusion and time decay through evaporation is proposed. The pheromone server can be embodied by a regular computer, a handheld device, or an embedded controller carried by a robot. Individual robots, and robotic swarms can be controlled this way. It is emphasized that the whole system, comprising the pheromone server, the communication system, and the robotic agents act like a neural network. This technique can lead to significant simplification of the task of controlling individual robots and robot swarms.

Keywords

RobotMobile robotComputer sciencePheromoneController (irrigation)Distributed computingHuman–computer interactionEmbedded systemArtificial intelligenceBiology

Related papers

Browse all SWARM papers