A Visual Servoing approach for road lane following with obstacle avoidance

This paper presents a local navigation strategy with obstacle avoidance applied to autonomous robotic automobiles in urban environments, based on the validation of a Visual Servoing controller in a Dynamic Window Approach. Typically, Visual Servoing applications do not consider velocity changes to stop the robot in danger situations or avoid obstacles, while performing the navigation task. However, in several urban conditions, these are elements that must be deal with to guarantee the safe movement of the car. As a solution for this problem, in this work a line following Visual Servoing controller will be used to perform road lane following tasks and its control outputs will be validated in an Image-Based Dynamic Window Approach. The final solution is a validation scheme for the Visual Servoing velocities which allows the obstacle avoidance, taking into account the car kinematics and some dynamics constraints. Experiments in simulation and with a full-sized car show the viability of the proposed methodology.