Analysis And Verification Of Six-DOF Manipulator Control Algorithm Based On Neural Network

Robot arm is widely used in resource collection,industrial production and life services and other fields.With the further study of the third generation of robotic arm,the intelligent level of the manipulator has been paid more and more attention.The control method with feedback has become the hotspot,and the non-calibrated visual servo control is one of them.Manipulator uncalibrated visual servo control refer to the control method which based on the image in the camera as a feedback to design the controller to guide the robot to complete the task,at the same time the camera's internal and external parameters are not calibrated.For uncalibrated visual servo control arm study began in the 1990 s.At present,there are algorithms based on image Jacobian matrix and algorithms based on other intelligent fitting.The two algorithms have their own advantages and disadvantages,and the Jacobi matrix is large and there is a singularity problem.The intelligent fitting algorithm is good in real time but the sample dependency is high,and how to satisfy both robustness and real-time performance is studied in recent years.Hot spots.In view of this problem,the main work of this paper is as follows:(1)The model of the manipulator and the visual mapping model are analyzed,and the simulation experiment platform is built according to the analysis.(2)The algorithm of uncalibrated visual servo control based on image Jacobi and the algorithm of uncalibrated visual servo control based on intelligent fitting algorithm are analyzed.In the simulation platform,two kinds of methods are simulated and verified.(3)The algorithm of uncalibrated visual servo control based on neural network is proposed and analyzed.Firstly,the monotonic relation between the image space and the Cartesian space is established by analyzing the visual model.The BP neural network is introduced to adjust the parameters of the controller to make the system deviation closer to the deviation in the real Cartesian space,so that the motion control is closer to the Cartesian The real state of space.Then,the Lagrangian dynamics equation and the general expression of the manipulator are used to prove the stability of the controller.Finally,the simulation experiment is carried out on the simulation platform.(4)In the physical experiment platform based dual-axis parallel binocular vision system and Denso six-degree-of-freedom manipulator,the algorithm of uncalibrated visual servo control based on image Jacobi and the algorithm of uncalibrated visual servo control based on intelligent fitting algorithm,and the algorithm proposed in this paper is verified experimentally.