Opportunistic Multi-sensor Fusion for Robust Navigation in Smart Environments

摘要

This paper presents the design of a navigation system for multiple autonomous robotic platforms. It performs multisensor fusion using a Monte Carlo Bayesian filter, and has been designed to maximize information acquisition. Apart from sensors equipped in the mobile platform, the system can dynamically integrate observations from friendly external sensing entities, increasing robustness and making it suitable for both indoor and outdoor operation. A multi-agent layer manages the information acquisition process, making it transparent for the core filtering solution. As a proof of concept, some preliminary results are presented over a real platform using the part of the system specialized in outdoor navigation.