Find in Library

Aiming at hysteresis changing with the load between the load torque and torsion angle of the harmonic reducer and a reduction of the conversion accuracy of the harmonic reducer resulting from this，a harmonic reducer with a memristor hysteresis model and an RBF neural network in parallel hybrid hysteresis model is proposed. The memristor model is improved to obtain a memristor hysteresis model，which is used to describe the basic change law of the hysteresis output of the harmonic reducer. The difference between the hysteresis model of the harmonic reducer and the memristor hysteresis model is compensated by the RBF neural network with nonlinear fitting ability. Experimental data verification results show that，compared with the memristor hysteresis model，the constructed hybrid hysteresis model can effectively describe the abrupt and non-smooth characteristics of the hysteresis of the harmonic reducer. The model validation mean square error is 0.06.