Autonomic Agents for Real Time UAV Mission Planning

This paper investigates agent-based modelling and design of a real time mission planning module for an Unmanned Aerial Vehicle (UAV). Agents can act on behalf of remote human operators, who impose mission objectives, performing autonomous or supervised behaviours. The planning module is in charge to deliberate the strategic tasks to achieve mission objectives and to satisfy environment constraints (area of operations, aircraft performances, etc.). We define a multi-objective model for implementing a goal-directed decision making support to the planning module. A formal model is used to infer from requirements real time constraints, by which mission planning is dynamically scheduled. Preliminary development and performance evaluation using embedded devices are discussed.