A Continuous Control Method for Task Stability of Visual Servo System

A robotic visual servo system using the sensor image as task target provides the possibility to expand the applications of industrial robots. However, the stability of the robot servo motion is not sufficient for industrial applications that have high precision requirements. In this paper, an adaptive visual servo algorithm based on second-order non-homogeneous linear differential equations (SNLDE) is proposed. By achieving a smooth transition during the change of robot joint velocities between two iterations of consecutive image tasks, this algorithm aims to improve the stability and reliability of the robotic system. Velocity changes of the robotic system under a traditional algorithm, first-order non-homogeneous linear differential equations (FNLDE) based algorithm and the SNLDE algorithm are compared in MATLAB, and the theoretical performance is verified by these tests. Subsequently, the stability of a robot under the three aforementioned algorithms is tested in a laboratory using six-axis industrial robots. The test results corroborate the effectiveness of the algorithm proposed in this paper.