Detection and tracking quadrotor using surf and feedback linearization sliding mode control

This paper presents a quadrotor detection and tracking system that uses a single camera and the Speeded Up Robust Features (SURF) algorithm to extract the image structure and estimate the target's position. The goal is to develop an independent system capable of detecting a followed object’s motion without prior knowledge of its trajectory, without requiring communication with the target, and without relying on external sensors. Feedback Linearization (FL) combined with sliding mode control is used to ensure target tracking. The control system avoids the complex nonlinear control solutions and the highly coupled dynamic behavior of the quadrotor. Nonlinear uncertain disturbances are overcome by using Time Delay Estimation (TDE) of the disturbance.