Study on Aerodynamic Parameter Estimation Method Based on Wavelet Neural Network and Modified PSO Algorithm

Abstract

Abstract This paper puts forward an aerodynamic parameter estimation method which combines the Wavelet Neural Network(WNN) with Modified Particle Swarm Optimization(MPSO) technique. This method directly and accurately constructs the relationship among flight parameters and aircraft aerodynamic parameters. Preliminary aerodynamic parameters and derivatives are derived from wind tunnel data. With strong nonlinear mapping ability, WNN is used to construct the relationship among Mach number, angle of attack, and rudder with longitudinal aerodynamic parameters. Then, the MPSO is used to estimate aerodynamic parameters based on the test-flight data. And WNN is retrained to amend the relationship among Mach number, angle of attack, and rudder with longitudinal aerodynamic parameters. Simulation verification indicates that MPSO has better estimation accuracy than Maximum Likelihood(ML) method. Comparison results of simulation experiments and flight-test data of a tactical missile show that simulated data based on estimated parameters matches with the flight-test data, which prove the effectiveness and validity of the trained WNN.