Application of neural network and spheric harmonics function to system error correction of electro-optical tracker system with X-Y gimbal

摘要

Correction of the system error of electro-optical tracker system could improve the measurement accuracy and guidance accuracy. Least square method has lower accuracy than spheric harmonics function method. Neural network could fit precisely complicated curves or curved faces and the system error of electro-optical tracker system with X-Y gimbal is locate on a curve composed of angle values. Analysis and simulation prove that BP neural network method improves the accuracy than spheric harmonics function method for about 8% and could have wider application scope.