Attitude control of UAV bicopter using adaptive LQG

This paper aims to design a controller that is able to maintain the stability of the unmanned aerial vehicle (UAV) bicopter attitude when carrying a payload. When the value of the payload inertia is in uncertainty, it is necessary to design a controller that can carry out the adaptation process. This paper proposes an Linear Quadratic Gaussian (LQG) adaptive controller to control the attitude of the bicopter with uncertain payload conditions. The proposed adaptive mechanism is a development of LQG control that can follow the response of the reference model. The success of LQG adaptive control is tested by providing uncertain payload parameters. The simulation results show that the LQG adaptive controller successfully overcomes the influence of inertial disturbances originating from the payload. There is a gain $\rho$ in the LQG adaptive mechanism, this gain is influenced by the parameter $\sigma$ which acts as a learning rate that produces a response to adapt to the response of the reference model. From the test results obtained when the value of $\sigma$ is enlarged there is an increased overshoot condition/value but the root mean square error (RMSE) value decreases. That means when the RMSE decreases, the response is getting closer to the model reference. To reduce the overshoot effect of increasing the value of $\sigma$, an improvement is made in the search for the gain value of $\rho$. From the test results, the value of $\sigma =1$ was chosen with the development of the gain equation $\rho$.