Estimation and Control for Autonomous UAV System to Neutralize unknown Aerial Maneuvering Target

Abstract

Autonomous capture of aerial non-cooperative target is a complex task requiring expertise from various fields like mechanical design, computer vision as well as estimation and control theory. The current work aims at presenting an estimation and control system as a part of an end to end solution for detecting, tracking and finally grabbing a dynamic aerial target. Estimation and control is typically challenging when the target under consideration is a maneuvering one. An Extended Kalman Filter is used to estimate the real time pose of the target with respect to the pursuer drone. This is followed by estimation of smooth trajectory which is fed to a gain scheduling based PD controller in order to engage head on with the target. Further, the accuracy of estimation algorithms and the ability of the proposed system to neutralize the target is demonstrated by means of simulation and experimental results.