Decision issues for multiple heterogeneous vehicles in uncertain environments

This paper focuses on on-going work in the “Action” project related to the decisional processes that are required to allow teams of cooperating heterogeneous vehicles to achieve target localization and tracking missions. Autonomy is made possible thanks to embedded functions such as data fusion, planning and supervision, organized within each vehicle so that they can cooperate together. Two essential characteristics are taken into account: the uncertainty on the environment models and on the vehicle action models (motions, perception and communication) and the constraints on the inter-vehicle communications, that cannot be permanently maintained. The paper presents the overall objectives of the Action project, analyzes the raised problems, and sketches the global approach chosen for the planning and supervision processes.