The precise control of the joint servo motor system depends on the mechanical parameters of the system. To address the limitations of traditional two-state extended sliding mode observer (ESMO)

which need prior friction coefficient calibration and exhibit structural complexity with strong coupling in multi-ESMO networks

this paper establishes a motion model equivalent to the motor side based on considering the transmission characteristics of the harmonic reducer and the characteristics of the system friction torque. By adding a new equation of state

the self-correction mechanism of the expanded sliding mode observer and sliding mode gain is designed to achieve the synchronous online identification of three parameters of the system

namely the moment of inertia

the viscous friction coefficient and the lump torque. Experimental studies show that the proposed method can effectively solve the problem that the convergence speed and identification accuracy cannot be taken into account together in traditional ESMO

and the total identification time of the three parameters can be shortened by at least 50%. When the parameter identification results are applied to the control of the joint servo motor system

the overdraft and adjustment time can be effectively reduced

with a reduction rate of up to 50%.