A new pose accuracy compensation method for parallel manipulators based on hybrid artificial neural network

The pose accuracy of parallel manipulators is one of the most important performance indices in advanced industrial applications. The modeling and estimation of geometrical parameter errors are no longer an issue because a large number of kinematic calibration techniques have been used to compensate pose accuracy of parallel manipulators in recent years. The modeling and identification of non-geometrical parameter errors are an extremely complicated technical procedure. A hybrid artificial neural network which involves a BP neural network, a RBF neural network and a control module is presented and used for compensating pose error caused by non-geometrical parameter errors of parallel manipulators. The control module serves as a linear mapping network which combines the outputs of the BP neural network and the RBF neural network to obtain the final pose accuracy compensation results. The pose accuracy compensation methods of the hybrid artificial neural network are established and discussed. Simulations and experiments were performed on a parallel manipulator. The feasibility and validity of the proposed pose accuracy compensation method based on the hybrid artificial neural network are verified through pose accuracy improvement and enhancement.

Automated summary

A pose accuracy compensation method based on a hybrid artificial neural network is proposed, which utilizes BP neural network, RBF neural network, and a control module to compensate for pose errors caused by non-geometrical parameter errors of parallel manipulators. Through simulations and experiments, the feasibility and validity of the proposed method are verified with improvements in pose accuracy.