MODELLING AND SIMULATION OF OCTOROTOR UAV (UNMANNED AERIAL VEHICLE)

• Main Author: HILAL TAWAB (NIM : 13206085); Pembimbing I : Prof. Dr. Ir. Bambang Riyanto Trilaksono; Pembi, TAUFIQ
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/17384
• Institution: Institut Teknologi Bandung

The development of UAV, recently, has been increased rapidly. The usage of UAV has been expanded to various areas, such as: military, farming, patrol, etc. The expansion of UAV can surely create a new challenges to the usage of UAV in a more difficult mission. Various designs have been proposed to answer those challenges. The design <br /> <br /> <br /> process of UAV must be supported by a good modeling method. In this case, we think that first principle approach is one of the cheapest, safest, and effective modeling method to obtain a good model representing the behavior of UAV. First principle approach is done by observing forces and moments that happened in UAV. By using second newton law and coriolis theorem, we can derive the non-linear equations which represents the behavior of UAV. Variable concerned in this UAV model will be three translation velocity in xyz coordinates, three angular velocity in xyz coordinates, and three angular position relative to inertial frame of reference in earth. In this modelling process, there are three kind of forces counted in the model, which are, <br /> <br /> <br /> gravitational forces, gyroscopic effects, and forces directly produced by the propellers. Those non-linear equations are linearized using taylor expansion and little perturbation method to produce linear system of octorotor. Each flying conditions represented by different linear system since the element of matrices A and matrices B will be different for different flying conditions. Non-linear model and linear model are simulated in Matlab, where the non-linear model is simulated using simulink block and the linear model is simulated using m-file. The results of both model will then be compared using three distinct inputs in two distinct flying conditions. The comparation results show that there are differences in few states because the linear model assumes that the octorotor is flying in stationary conditions with little perturbation and zero perturbation derivatives. In this case, we can conclude <br /> <br /> <br /> that the non-linear model is better in representing the dynamics of octorotor than the linear ones. The experiment data are compared by the model. The similarity of the model response and the experiment data is calculated using best-fit method. The comparison between filtered experiment data and model is also done. In this case, the filter used is <br /> <br /> <br /> butterworth low-pass filter with a cut-off frequency of 12 rad/s. The results of best-fit show that the model response has been represented the behavior of ocotorotor quite well, <br /> <br /> <br /> which the highest best-fit value is 75.37% and the lowest value is 30.9%