Extended Kalman Filter Based Fuzzy Decentralized Sliding-Mode Trajectory Tracking Control of Quadrotor Unmanned Aerial Vehicles

• Main Authors: Zih-Siang Lin, 林梓翔
• Other Authors: Chih-Lyang Hwang
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/8y4mua

碩士 === 國立臺灣科技大學 === 電機工程系 === 104 === Due to advantages and disadvantages of inertia navigation system (INS) and global positioning system (GPS), it is not suitable for the only use of GPS or INS for the measured velocity and position of outdoor QUAV. Furthermore, the computation of the rotation described by Euler angle is inefficient; its description sometimes possesses the singularity problem. Hence, it is not suitable for the embedded single board computer to calculate the corresponding signals using the Euler angle based rotation description. In this thesis, the nonlinear mathematical model of sensors including GPS and INS is first established. Its linearized model around the GPS signal is constructed for the discrete version of extended Kalman filter (EKF). The estimated position and velocity from EKF are employed to correct the position and velocity of INS such that the controller design is more effective. The compared performances between GPS and INS, and GPS-aided INS confirm the effectiveness of the estimated signal through GPS-aided INS. Subsequently, the EKCF-based fuzzy decentralized path tracking control (FDPTC) is applied for the path tracking of an outdoor QUAV. The proposed control does not need the mathematical dynamic model of QUAV, it only needs its input/output data to construct the fuzzy rule table. After that, three factors and the coefficients of sliding surface for each subsystem are tuned to obtain the satisfactory control performance. Finally, the compared experiments of circular path confirm the effectiveness and robustness of the proposed method.