Analyzing the Effectiveness of State Estimation Strategies for Quadrotor Attitude Models

Quadrotors have become extremely popular in a number of areas with their ability to control their attitude. There are many control methodologies of state-feedback that are created to optimize nonlinear multivariable control. This paper presents a state estimation scheme founded on subspace identification (SI) in quadrotor attitude control, which is useful to work in the conditions of no explicit model and with noisy signal conditions. The main objective is to come up with an observer-based discrete state feedback control system that would satisfy the desired attitude control performance objectives. In order to have the correct modeling and estimation of state, a pseudo-random binary sequence (PRBS) is produced as the reference input, which is optimized according to the settling time and operating frequency identified with the help of closed-loop experiments. The given SI solution is aimed at overcoming the difficulty of estimating the yaw-axis state of the quadrotor attitude model with interference and measurement noise. A control framework (CF) model of high order is first being identified by the 4S identification technique and model order reduction by the balanced stochastic model truncation (BSMT) technique. Lastly, the similarity transform is employed to determine the relationship between the actual quadrotor model and the computer-based CF model. The analysis of the control loop performance and its validation are mainly in the yaw-axis attitude control loop.