Rotary Unmanned Aerial Vehicles (RUAVs) suffer in average Northern Irish winters due to heavy wind preventing vital tasks from being performed in the economy, such as search and rescue or civil engineering observations. This work provides enhanced stability of RUAVs under wind disturbances by using metaheuristic algorithms to select optimal controller gains. Previous work demonstrated how Particle Swarm Optimization can be used to tune optimal controllers; this work uses a machine learning algorithm (Genetic Algorithm) to tune the controller. Simulations carried out on Full State Feedback, Full State Compensator and Linear Quadratic Gaussian controllers tuned by a variety of techniques revealed that the Genetic Algorithm outperformed conventional manual tuning by 20% and Particle Swarm Optimization by 17% in performance measured in settling time. The proposed method tunes the feedback gains and Kalman filter by Genetic Algorithm, which outperforms the manually tuned conventional schemes and “GA-Hybrid” approach. The conditions required to employ Reinforcement Learning as an alternative method for RUAV stabilization in future scope is also explored.
Bibliographical noteFunding: This research received no external funding.
- optimal controller
- particle swarm optimization
- genetic algorithm
- hybrid control