Estimation of nonlinear airship parameters using modular neural network

The prime objective of this work is to estimate stability and control derivatives of an airship. The complete, nonlinear mathematical model of aerial vehicles has its aero model as a nonlinear function of angle of attack. This along with the necessity for an exhaustive dataset complicates the estimation procedure. In this work, data are generated by simulating the mathematical model of airship for different trim conditions obtained from continuation analysis. A modular neural network is then trained using back-propagation and Adam optimization algorithm for each aerodynamic coefficient separately. The estimated nonlinear airship parameters are found to be consistent with the DATCOM parameters which were used for open-loop simulation in data generation phase. This validates the proposed methodology and could be extended to estimate airship parameters from real flight data.