IMPLEMENTASI SISTEM KENDALI ADAPTIF SELF-TUNING REGULATOR UNTUK KENDALI SIKAP WAHANA QUADCOPTER

• Main Author: Dayu Alifany Hardy, Damar
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/43536
• Institution: Institut Teknologi Bandung

This thesis discussed about the design of an adaptive controller using direct self-tuning regulator to be used for quadcopter air vehicle. Common controller that used by many quadcopters is constant gain PID. This linear controller need to be retuned for any parameter change within the quadcopter. The ability of adaptive controller to detect any change in quadcopter dynamic system is generated by the parameters that estimated using Recursive Least Square algorithm. This RLS algorithm also take advantage from using variable forgetting factor to estimate any new parameter change while the simulation is running. These estimated parameters will be used for calculating command signal using direct STR controller. The direct STR itself uses a method called pole-placement to control the respon of quadcopter model. Generally, the working principle of pole-placement method is by changing the closed-loop poles to a value that already defined by user so that the roots will be the same as the reference. Phenomenon where quadcopter responses act like the reference is called model –following. Flight control system that used are attitude hold and altitude hold. Besides, RLS algorithm and direct STR algorithm is also embedded to an Arduino microcontroller so that HIL Simulation can be executed for SISO case. Simulation result shows that RLS algorithm can produce a good result in estimating parameter for a linear model. But there are some difficulties to converge when estimating a nonlinear model. The direct STR controller successfully generate a proper command signal so that quacopter’s respon already the same to what it is referenced for SIL simulation scheme. While for HIL simulation scheme, it happened that direct STR couldn’t really calculate a proper command signal due to some factors. However, flight control system for altitude hold and attitude hold can be done good enough even there was a change in quadcopter mass during the simulation