Model-Free Visual Servo Regulation of Wheeled Mobile Robots Under Dynamic Visual Targets

In this paper, a new method is proposed for wheeled mobile robot visual servoing using model-free visual targets under dynamic scenes. In existing methods, feature points are generally supposed unmovable in visual servoing stabilization process. For dynamic scenes, A monitor camera must be added to the workspace in order to complete visual servoing tasks. On this basis, in the first stage, the ratio between depth information is calculated by using the geometric relationship between feature images and points. Subsequently, relative relationship between various coordinate systems can be calculated via homography decomposition techniques and coordinate transformation. Finally, an adaptive controller is used to drive mobile robot to achieve the desired pose. The simulation results show the viability of the strategy.