Autonomous UCAV Strike Missions Using Behavior Control Lyapunov Functions

摘要

An autonomous Unmanned Combat Aerial Vehicle (UCAV) carrying out a surveillance or strike mission must be able to handle situations where the dierent mission objectives are in conflict and a tradeo must be made, e.g. when the time of arrival is in conflict with the prescribed safety distance to an enemy surface to air missile (SAM) site. This paper describes a framework called Behavior Control Lyapunov Functions (BCLF), to handle such tradeos. The framework combines the natural idea of dierent control behaviors for dierent mission objectives, suggested in the Behavior Based robotics approach, with the mathematical transparency of Control Lyapunov Functions (CLF) from control theory. First, each behavior is represented by a scalar function with certain CLF-like properties, describing to what extent that mission objective is satisfied. The operator then edits a priority table reflecting the order of importance between dierent objectives, as well as dierent levels of satisfaction. Based on the table and the current levels of satisfaction the algorithm decides which objectives should be focused on right now, and which should currently be ignored. Finally, the current high priority objectives are transformed into recommended subsets of the available control choices, and passed to the controller. The paper is concluded with simulation examples illustrating the approach.