Robust Nonlinear Autopilot Design Based on Feedback Linearization and Extended State Observers for Interceptors

The conventional methods for designing autopilots deal conservatively with uncertainties such that the autopilot should be slowed down to the extent in which the system can maintain its desired performance in the presence of a certain uncertainty. However, there will be no guarantee of good performance in the presence of all uncertainties while the capabilities of the system will not be used properly. In the proposed method, using the extended state observer, the lumped uncertainties along with the system’s states will be estimated, and using feedback linearization approach, along with the nonlinear terms will be removed from the control loop. What remains is a simple linear system that is easier to control. In this method, uncertainties is treated without conservatism and consequently performance is improved in different conditions. Finally, for comparing the results of proposed method with linear autopilot a 1000 times Mont Carlo running with specific condition has been used and is shown that the proposed method is better than the conventional method in performance.