In this paper we introduce a novel cloud robotics architecture that provides different functionalities to support enhanced coordination of groups of Automated Guided Vehicles (AGVs) used for industrial logistics. In particular, we define a cooperative data fusion system that, gathering data from different sensing sources, provides a constantly updated global live view of the industrial environment, for coordinating the motion of the AGVs in an optimized manner. In fact, local sensing capabilities are complemented with global information, thus extending the field of view of each AGV. This knowledge extension allows to support a cooperative and flexible global route assignment and local path planning in order to avoid congestion zones, obstacles reported in the global live view map and deal with unexpected obstacles in the current path. The proposed methodology is validated in a real industrial environment, allowing an AGV to safely perform an obstacle avoidance procedure.