The development of modern manufacturing requires a higher speed and accuracy of coordinate measuring machines (CMM). The dynamic error is the main factor affecting the measurement accuracy at high-speed. The dynamic error modeling and estimation are the basis of dynamic error correcting. This paper applies generalize regression neural network (GRNN) to establish and estimate dynamic error model. Compared with BP neural network (BPNN), GRNN has less parameters, only one smoothing factor parameter should to be adjusted. So that it can predict the network faster and with greater computing advantage. The running speed of CMM axis is set through software. Let it running for the X axis motion. The values of the grating and the dual frequency laser interferometer are gained synchronously at the same measure point. The difference between the two values is the real-time dynamic measurement error. The 150 values are collected. The first 100 values of the error sequence are used as training data to establish GRNN model, and the next 50 values are used to test the estimation results. When the smooth factor is set at 0.5, the estimation of GRNN training data is better.The simulation with the experimental data shows that GRNN method obtains better error estimation accuracy and higher computing speed compared with BPNN. GRNN can be applied to dynamic error estimation of CMM under certain conditions.
[1]
Phj Piet Schellekens,et al.
Development of a fast mechanical probe for coordinate measuring machines
,
1998
.
[2]
Georgina Stegmayer,et al.
Estimation of the Particle Size Distribution of a Latex using a General Regression Neural Network
,
2008,
IFIP AI.
[3]
Hongtao Wang,et al.
Study of coordinate measuring machines synthetic dynamic error under different positions and speeds based on dual linear returns
,
2006,
International Symposium on Precision Mechanical Measurements.
[4]
Hilmi Berk Celikoglu,et al.
Public transportation trip flow modeling with generalized regression neural networks
,
2007,
Adv. Eng. Softw..
[5]
Hikmet Kerem Cigizoglu,et al.
Generalized regression neural network in modelling river sediment yield
,
2006,
Adv. Eng. Softw..
[6]
Raouf N. Gorgui-Naguib,et al.
A general regression neural network analysis of prognostic markers in prostate cancer
,
1998,
Neurocomputing.
[7]
Peter Wolschann,et al.
Comparison between two types of Artificial Neural Networks used for validation of pharmaceutical processes
,
2009
.