LQG-LTR controller design to single-link rotary flexible joint

In this thesis, a methodology for a linear control structure that can be used in the presence of inadequate measurement conditions for a single-linked flexible manipulator is presented. The Linear Quadratic Regulator (LQR) structure is extended with integral action. A Linear Quadratic Gaussian (LQG) control system with integral (LQG-I) support is designed on the extended system. A Kalman-based Linear Quadratic Estimator (LQE) is used to estimate unmeasured states. The Loop Transfer Recovery method (LTR) is applied to increase the stability of the LQG, approximating the LQG controller to the LQR. The LQR structure was extended with an integral effect, and the performance of an LQG-I controller supported by LQE and LTR was analyzed using a dynamic model, an experimental single-joint rotating flexible joint. The model was analyzed in the frequency domain, and LTR analysis was performed with the LQG input defined. The system's response at different inputs was analyzed using the specified parameters. The LQR controller, which can measure all states, was compared with the LQG controller, which can measure only one state and estimate the other states.