engleski

Decentralized Energy Sharing Protocol using TÆMS Framework and Coalition-based Metaheuristic for Heterogeneous Robotic Systems

In this paper we describe a decentralized protocol that is applied for scheduling and management of energy sharing in heterogeneous systems. Particular system we have in mind comprises of agents with different mobility, energy harvesting and energy sharing capabilities herein called aPads, aFish and aMussels. aPads are surface vehicles that harvest energy and are able to transmit it to both aFishes and aMussels. aFish are underwater vehicles that can transmit energy to aMussels or to other aFish, and receive energy from aPads. aMussels, that can either be attached to the sea-ground or float on the surface, can only receive energy, either from aFish or aPads. For such a system to be able to function for an extended period of time, a decentralized energy management protocol should be designed, whose purpose is to schedule and coordinate charging of agents over time. These decisions depend on the current battery levels of different agents, as well as their positions, and should be done in a decentralized manner. In this paper we apply an energy sharing protocol based on TÆMS framework, originally applied to planning and coordination in aerial- ground robotics systems for search and rescue missions. We refine the obtained plans using the coalition- based metaheuristic for vehicle routing problems. We show execution schedules achieved using the proposed protocol and highlight its advantages and flaws when applied to our underwater scenario.

hierarchical decision making ; decentralized systems ; scheduling algorithms ; distributed artificial intelligence

nije evidentirano