Application And Research Of PSO-GA-BP Neural Network In Robot Uncalibrated System

According to the error information of the target features in the image space,the uncalibrated visual servo system maps the plane image error to the three-dimensional space through the visual mapping model,and guides the manipulator to realize the spatial positioning.Uncalibrated visual servo system has a strong adaptability in uncertain work scenes and complex tasks.In this paper,by analyzing the current problems of uncalibrated visual servo control system,a visual servo control system based on particle swarm genetic algorithm optimized BP neural network(PSO-GA-BP)is designed.In this paper,image preprocessing is carried out for the collected image,Canny operator is used to detect the edge of the feature,and the random Hough transform(RHT)ellipse detection algorithm with the constraint of the feature string is used to extract the image positioning feature.Before the boundary fitting,the validity of detecting the circular feature center with the constraint of ellipse power is used,which reduces the cumulative number of votes and improves the efficiency and accuracy of feature detection.The vision servo scheme of eye in hand structure is constructed,the image Jacobian matrix is analyzed and deduced,and the differential mapping relationship between manipulator motion space and image feature space is established.Aiming at the problems of BP neural network,such as slow convergence speed,long training time and easy to fall into local extremum,the PSO-GA algorithm proposed in this paper is used to optimize the initial weight and threshold value of BP neural network,and to carry out crossover and mutation operations on particles during particle swarm iteration,so as to maintain the diversity of population,prevent population from falling into local extremum and improve the search ability of population.In this paper,a visual servo controller based on PSO-GA-BP neural network is designed and established.The experimental results show that the optimized algorithm has high efficiency.Compared with the traditional Jacobian matrix control algorithm,it can make the robot end effector reach the expected position in a shorter time.The control system has good convergence speed and control accuracy,and is stable and effective.