Adaptive Kalman Filtering in LQG Control of an Uncrewed Helicopter

This study presents the design of an automatic flight control system that combines a Linear Quadratic Regulator (LQR) with an adaptive Kalman filter state observer, resulting in an adaptive LQG-type controller. The proposed approach was demonstrated using a high-fidelity simulation model of an unmanned helicopter developed at the Warsaw University of Technology. The controller parameters were tuned using the Particle Swarm Optimization (PSO) algorithm to minimize the control error while accounting for onboard actuator limitations. The adaptive feature of the Kalman filter lies in the online adjustment of the process noise covariance matrixQduring rapid maneuvers, which effectively reduces the state estimation errors under highly dynamic flight conditions. The main objective of this research is to present a complete framework for synthesizing automatic flight control systems for unmanned rotorcraft based on a simplified linear model of vehicle dynamics. The integration of PSO-based control tuning with adaptive adjustment of the process noise covariance matrix provides a unified methodology for developing efficient and robust flight control systems.