Intelligent Parameter Identification for Robot Servo Controller Based on Improved Integration Method

With the rise of smart robots in the field of industrial automation, the motion control theory of the robot servo controller has become a research hotspot. The parameter mismatch of the controller will reduce the efficiency of the equipment and damage the equipment in serious cases. Compared to other parameters of servo controllers, the moment of inertia and friction viscous coefficient have a significant effect on the dynamic performance in motion control; furthermore, accurate real-time identification is essential for servo controller design. An improved integration method is proposed that increases the sampling period by redefining the update condition in this paper; it then expands the applied range of the classical method that is more suitable for the working characteristics of a robot servo controller and reducesthe speed quantization error generated by the encoder. Then, an optimization approach using the incremental probabilistic neural network with improved Gravitational Search Algorithm (IGSA-IPNN) is proposed to filter the speed error by a nonlinear process and provide more precise input for parameter identification. The identified inertia and friction coefficient areused for the PI parameter self-tuning of the speed loop. The experiments prove that the validity of the proposed method and, compared to the classical method, it is more accurate, stable and suitable for the robot servo controller.