Comparison of bayesian regularization and Optimal Brain Damage methods in optimization of neural estimators for two-mass drive system

In this paper an implementation of optimized neural networks for state variable estimation of the drive system with elastic joints is presented. The signals estimated by neural networks are used in the control structure with state-space controller and additional feedbacks from the shaft torque and the load speed. High quality of estimation is very important for correct operation of the closed loop system. The precision of state variables estimation depends on generalization properties of neural networks. Short review of optimization methods of neural networks is presented. Two techniques typical for regularization and pruning method are described and tested in details: Bayesian regularization and Optimal Brain Damage. Simulation results show good precision of both optimized neural estimators for a wide range of changes of the load speed and load torque, not only for nominal but also changed parameters of the drive system. The experimental results are also shown and a high quality of estimation is obtained in a laboratory set-up.