Anfis Based Longitudinal Controller Design of a Fixed Wing UAV

Ulus Ş. , Eski İ.

5th. International Conference on Engineering and Natural Sciences (ICENS 2019), Praha, Czech Republic, 12 - 16 June 2019, pp.310-317

  • Publication Type: Conference Paper / Full Text
  • City: Praha
  • Country: Czech Republic
  • Page Numbers: pp.310-317


In this study, adaptive neuro-fuzzy inference system called ANFIS structure was investigated for longitudinal control of a fixed wing unmanned aerial vehicle (UAV). ANFIS control is one of the modern control techniques which is based on artificial neural networks (ANN) and the structure of the system is based on Takagi-Sugeno fuzzy inference system. In this conference paper, ANFIS control method was studied in order to design a longitudinal controller architecture of a fixed wing UAV. Reference model and aerodynamic parameter values of the UAV were obtained from the literature for this specific UAV model. According to the simulation data taken from PID controller structure in MATLAB®, ANFIS training and testing data were obtained and trained data were used for longitudinal aircraft dynamics such as pitch attitude hold characteristics. Hybrid and back-propagation optimization algorithms were applied to have optimum training performance. ANFIS structure was also simulated together with PI and PID controllers in Simulink. In this analysis, results have been satisfying in terms studies of the fixed wing UAV control and ANFIS and PID hybrid methods had more suitable results