Visual Servoing Control Algorithms for a Crop Row Navigation of a Four-Wheel Steering Robot

In this paper, a comparison between image-based visual servoing and position-based visual servoing methods for a crop row path-following four-wheel drive four-wheel steering robot is presented. A cascade control system with an external loop for the visual servoing kinematic control and an inner loop for the dynamic robot control is considered. The image processing time delay is also considered in the proposed algorithm to ensure the system convergence, along with a filtering procedure to limit the robot's maximum accelerations and a prediction algorithm to increase the control robustness against image loss. The results obtained using the proposed methods are validated and compared through simulations.