Hybrid localization solutions for robotic logistic applications

This paper presents some results concerning rovers localization, achieved within the framework of the Italian Regional project MACP4Log. This project investigates the coordination of a team of rovers that delivers assistance services in the context of logistic spaces. The issue of single rover localization is a service prerequisite which is faced by different approaches. First a WiFi-based localization technique is examined: it involves building a radio map of the Received Signal Strength (RSS) in the considered environment and, afterwards, setting the problem of localization as an optimization problem: given a RSS, the function is reversed to get the estimated position of the rover. Finally the proposed WiFi technique is integrated with a particle filter algorithm following both a "loose" and a "tight" coupling approaches, integrating radio information at two different levels. Experimental results in monodimensional and bidimensional environments validate these approaches and allow a comparative analysis among them.