PLS -WNN algorithm and its applications in aerodynamic parameters regression estimate

摘要

An aerodynamic parameters regression estimate method based on Wavelet Neural Network by Partial Least Square feature extraction is proposed. This method can overcome problems of data noise and multiple correlations among parameters, accurately describe the dynamic characteristics of flight vehicle. Firstly, using Partial Least Square extracts basic feature of training samples in flight data. The second, aerodynamic parameters are regression estimated based on Wavelet Neural Network by using basic feature extracted. Finally, the method proved by experiment is effective and feasible to be used for flight vehicle aerodynamic parameters regression estimate.