Research On Intelligent Control Algorithm Of Robot Manipulator Based On Visual Servoing

With the rapid development of industry and the demand of modern society,robot manipulators have become more and more widely used.Robotic visual servoing system has become one of the most important directions for the research of modern intelligent robots.Since robotic visual servoing system involves the dynamics and kinematics of manipulators,image processing,machine learning,automation control and so on,which is a very challenging research topic.The image based visual servoing isn't dependent on system parameters of camera and manipulator,therefore,the image-based uncalibrated visual servoing method has become the mainstream of current research.Intelligent algorithms were combined with image-based visual servoing system in this paper,focusing on the combination of intelligent control technology and image features,to construct a better visual servoing intelligent control model and manipulator trajectory planning method.The visual servoing model based on image moments and the extreme learning machine model have been studied emphatically,which solved the problem of the rotation angle of the manipulator end-effector on the x-axis and y-axis when the image moments are used and the optimal trajectory planning problem of the manipulator of fixed-position movement of the least energy consumption.In general,the main work contents and research results of this paper are summarized as follows:(1)For the problem that it is easy to fall into local optimum in original Firefly Algorithm and its convergence is slow at late period.In this paper,a firefly algorithm based on adaptive inertia weight and individual mutation is proposed to improve the optimization ability and node convergence speed of the Firefly Algorithm.Then the improved firefly optimization algorithm is used to optimize the obtained and optimized input weights and biases of the hidden layer in extreme learning machine to obtain a stable IFA-ELM model.Afterwards,the IFA-ELM model is used to solve the invisible non-linear mapping relationship between the invariable image moment combination feature and the x-axis and y-axis rotation angle of the manipulator end effector in the visual servo system,when the image moment is used as the image feature.The simulation experiment results show that the improved extreme learning machine model is faster than the traditional neural network algorithm,and the estimation error is less than 0.25 degrees.The proposed visual servoing algorithm has better performance and stability.(2)In order to overcome the disadvantages of poor accuracy and poor robustness,it is due to the fact that the input weights and biases of the hidden layers in On-Line Sequential Extreme Learning Machine(OSELM)are randomly acquired,therefore,the Grey Wolf Optimizer(GWO)is used to optimize the input weights and biases of the hidden layers of OSELM to obtain the GWO-OSELM model,which is used to improve prediction accuracy of OSELM.We apply it to visual servoing system based on image moment feature to solve the invisible non-linear mapping relationship between the invariable image moment combination feature and the x-axis and y-axis rotation angle of the manipulator end effector.The simulation results show that the GWO-OSELM model has good generalization performance and stability,and has good prediction accuracy on regression problem,similarly,it can keep the robot's path stability in the target crawling motion,with strong robustness and stable performance in IBVS based on image moments.(3)In order to solve problem that the Grey Wolf Optimizer is easy to fall into local optimum,this paper designs a weighting factor that can dynamically change with the optimization process to improve the search speed and optimization ability of the Grey Wolf Optimizer.Then it uses the cubic spline interpolation method combined with the improved gray wolf optimization algorithm to plan the ideal trajectory for the robot joint space.The simulation results show that the cubic spline interpolation algorithm based on improved grey wolf optimization algorithm can generate smooth trajectories,and plan along the ideal trajectory to ensure that the energy consumed by the system during the movement is the minimum.At the same time,the algorithm also has the advantages of being simple and easy to implement.