Vision-based robust localization for vehicles

A popular technique to localise a vehicle on the lane is by using a camera mounted on the chassis pointing headway. This approach requires: an image processing algorithm to extract the relevant information from the scene and an estimation algorithm to estimate the position of the lane boundary. In this paper, we address both problems simultaneously by defining a robust estimation algorithm that simplifies the image processing and enables localization of the vehicle in the 3D space. The accuracy of the proposed solution is evaluated by simulations and experiments.