Multi-object tracking with distributed sensing

Abstract

One of the main research goals on distributed autonomous agents in a Multi-Agent System is the development of mechanisms to form a better world model using information merging from different agents. In this paper, we present a solution for robust online and real-time multiple object tracking in a multi-agent system using information gathered by various agents over time, using COP-KMeans for clustering and Kalman Filtering for object state estimation. The proposed solution was implemented on a real robotic soccer team and evaluated in the RoboCup Middle-Size League competitions. The robotic soccer was presented as one possible application for the ideas expressed on this paper.