Decentralized Constraint Optimization in Composite Observation Task Allocation to Mobile Sensor Agents

Abstract

Cooperative severance and observation by autonomous multiple mobile sensors have been studied for wide area monitoring, disaster response, and exploration in unsafe zones. In practical situations, sensor agents might be required to perform various composite tasks. To integrate them, the general representation of problems and decentralized solution methods for different requirements are necessary. The distributed constraint optimization problem has been studied as a general and fundamental combinational optimization problem in multiagent systems. Although several studies have applied this approach to sensor networks and teams of mobile sensors, opportunities also exit to apply it to manage composite tasks and utilize decentralized protocols in the subtasks in several layers of observation systems. As a case study, we address a cooperative observation system consisting of mobile sensor agents that temporally observe unsafe zones on a floor or a field with obstacles, where the basis of the tasks is the division of observation areas for the agents. We also allocate several tasks with high priority to several agents. We applied a decentralized constraint optimization method to the cooperation for both task allocation and the division of an observation area and experimentally verified our proposed approach in a simulated environment.