Abstract:  Aiming at the difficult control problem of assembly manipulator with shallow depth insertion, a hidden Markov based shallow depth insertion control strategy for manipulator was proposed. Firstly, the shallow depth insertion operation was analyzed, and the force control parameters were obtained by collecting manual operation data. Secondly, based on the measured manual operation data, the time series data of force control parameters were identified, and the manual operation process was modeled by using damping control law. Then, the hidden Markov model was used to model the time series data of the force control parameters, so as to obtain the control strategy suitable for the shallow depth insertion of the manipulator. The strategy was expressed by the symbolic output probability of the hidden Markov model in each state. Finally, UR10 6DOF robotic arm was used to conduct shallow depth insertion experiments in various application scenarios. The results show that mechanical arm can accurately obtain damping control parameters of the process of manual operation, get hidden Markov relative to the mean distribution of the standard deviation of the control parameters, 
thus realizing its application to various scenarios, which provides an effective theoretical basis for the autonomous execution of the shallowdepth insertion assembly of the mechanical arm.

Key words:  assembly of mechanical parts; robotic arm automated assembly; control strategy;shallow depth insertion; hidden Markov process