Intelligent visual servoing for nonholonomic mobile robots

Abstract

In this work the authors present a visual servoing approach based on particle swarm optimization (PSO-PBVS). The PSO-PBVS algorithm overcomes the traditional PBVS approach by keeping the target features inside the image plane during the servoing. The PSO-PBVS approach allows to define the 3D trajectory in the task space, in the same way as the PBVS approach. In addition, an intelligent control is used to estimate the currents for each motor, and ensure that the motors provide the desired velocities.