|Communications on Applied Electronics
|Foundation of Computer Science (FCS), NY, USA
|Volume 7 - Number 5
|Year of Publication: 2017
|Authors: Dominic Asamoah, Samuel Winful, Stephen Opoku Oppong
Dominic Asamoah, Samuel Winful, Stephen Opoku Oppong . Intelligent Vehicular Traffic Light Control using Hidden Markov Model. Communications on Applied Electronics. 7, 5 ( Aug 2017), 12-20. DOI=10.5120/cae2017652668
Traffic management continues to remains a major problem in big cities. Allocating limited resources, i.e. roads, to an increasing number of users with individual needs and objectives, turns out to be a highly complex in most cases. This research uses Hidden Markov Model (HMM) as a component with unsupervised clustering scheme to determine the traffic situation of a road in a traffic video. An unsupervised clustering algorithm called Autoclass is applied to obtain the traffic density state (free, normal and congested) on motion features which are extracted from each frame. The three HMM models are constructed for each traffic state with each cluster corresponding to a state in the HMM. The result show that this method can handle varying illumination and classify traffic density in a (Region of Interest) ROI of a traffic video.