This paper discusses the basic principles of fuzzy neural network and the method of GPS height conversion,and implements of the GPS height conversion by fuzzy neural network method.

The basic principle of RBF network is discussed,and the GPS height conversion by RBF method is implemerted.

In this paper, the fifing method which stimulates the Geoid using quadric polynomial, BP network and RBF network are provided for the GPS height conversion.

The application of support vector machine method in GPS height transformation;

The key issue of GPS height transformation is the calculation of the height anomaly.

Based on the introduction of calculation method about GPS normal height,this paper takes the actual data of a surveying area as an example,and programs with Matlab to carry out height transformation.

The paper discusses elevation conversion of GPS geodetic elevation on GPSurvey software with regularity fitting mathematics face.

Combined with the application of GPS control net in a large-scale water conservancy project in Guangxi,introduction and analysis was made on the elevation conversion method of transforming GPS ellipsoidal height into normal height.

In addition, in view of the problem of the　elevation value rectified, we present the methods of zero-value-line and the　field identification.

Studying of new method for GPS height transformation

The approach and mathematic model of GPS height transformation,and the application of remove-store approach based on Earth s gravity field model in GPS height transformation are researched.

GPS height conversion based on Simulated Annealing BP Neural Network Model;

The GPS height conversion model based on the BP neural network is designed.

GPS height conversion can achieve high precise with the arithmetic of BP neural network,but there still are some problems,such as how to determinate the number of layers for a network and the number of neu- rons for each layer,and so on.