Research On Filtering Method Of Robotic Kinematics Systems Based On Wavelet Analysis

The robotic assisted minimally invasive surgery (MIS) has the advantages of small trauma, less blood and pain, and quick recovery. Robots can implement complicated and accurate surgery, develop surgeons’operation ability, so the robotic assisted microsurgery has gained wide recognition in the surgery field, there has been great process and success in the study of this field home and abroad. Even so, for some surgeries that need non-teleoperation and directive operation, because of the physiological hand tremor in the robot system of MIS, there are certain deviations between the actual input action of operators and the desired input action, which will impact the surgery quality and lead to failure of surgery. Most of the existing methods for tremor filtering weaken and linearize the tremor signal, or identify the tremor only in time and frequency domains nonlinearly, these approaches cannot handle the fuzzy uncertain information coupled with tremor signal, and have large limitation in describe the characteristics of tremor.This paper integrated the features of MIS and hand tremor, studied a master-slave robotic system for microsurgery, and proposed a novel adaptive tremor filter, the optimization algorithm for parameters of the filter has been developed.In the first place, considering the features of robots for MIS, this paper introduced the related kinematics theory about master-slave robot for microsurgery, built kinematic model for the master-slave manipulators, and designed a modularized master-slave robotic system for MIS.Secondly, in light of the complicated features of tremor, this paper proposed a three-domain fuzzy wavelet network (3DFWN) filter for tremor suppression on the basis of analyzing adaptive filtering theory, the physical model, mathematical model and algorithm implementation of the filter were elaborated.Then, to make up the deficiency of optimization algorithms for neural network, this paper introduced particle swarm optimization (PSO) algorithm, on the basis of that, some improvements has been made by using fuzzy theory. The improved PSO algorithm was applied to optimize the structure parameters of the tremor filter. In the end, to validate the proposed adaptive3DFWN filter and the improved PSO algorithm, this paper carried out a series of comparative experiments based on MATLAB platform using different fitted tremor signal.