Application of Neural Network Parameter Identificationin Driving System With Sensor Characteristic

The paper introduces a method of linear system parameter identification based on Hopfield Neural Network. This method assumed that the input of Neural Network was system state variables which were tested and delayed by sensor. Full condition of identification was deduced, and this method was applied to identification of revolve value J, resistance R_Ω and load torque T_L of squirrel cage motor driving system. Simulation results showed that this method can make the identification results converqe in true value even if the system condition is deteriorated by not correct parameters of the controller.