Fast lane tracking for autonomous urban driving using hidden Markov models and multiresolution Hough transform

Purpose – Lane tracking is one of the most important processes for autonomous vehicles because the navigable region usually stands between the lanes, especially in urban environments. A robust lane tracking method is also required for reducing the effect of the noise and the required processing time. The purpose of this paper is to present a new lane tracking method. Design/methodology/approach – A new lane tracking method is presented which uses a partitioning technique for obtaining multiresolution Hough transform of the acquired vision data where Hough transform is one of the most popular algorithms for lane detection. After the detection process, for tracking the detected lanes, a hidden Markov model (HMM) based method is proposed. Findings – The results of the proposed approach show that the partitioned Hough transformation reduces the effect of noise and provides robust lane tracking. In addition, the acquired lanes are successfully tracked by using the designed HMM. Originality/value – This paper provides a fast lane tracking system which can be integrated with an autonomous vehicle or a driver assistance system.