This paper presents a method for modelling relationship between road segments using feed forward back-propagation neural networks. Unlike most previous papers that focus on travel time estimation of a road based on its traffic information, we proposed the Neighbouring Link Inference Method (NLIM) that can infer travel time of a road segment (link) from travel time its neighbouring segments. It is valuable for links which do not have recent traffic information. The proposed method learns the relationship between travel time of a link and traffic parameters of its nearby links based on sparse historical travel time data. A travel time data outlier detection based on Gaussian mixture model is also proposed in order to reduce the noise of data before they are applied to build NLIM. Results show that the proposed method is capable of estimating the travel time on all traffic link categories. 75% of models can produce travel time data with mean absolute percentage error less than 22%. The proposed method performs better on major than minor links. Performance of the proposed method always dominates performance of traditional methods such as statistic-based and linear least square estimate methods.

Travel time is a basic measure based on which intelligent transportation systems such as traveller information systems, traffic management systems, public transportation systems are developed. Although many methodologies have been proposed, they have not yet adequately solved many challenges associated with travel time, in particular, travel time estimation for all links in a large and dynamic urban traffic network is still an open problem that needs addressing. Typically focus is placed on major roads such as motorways and main city arteries but there is an increasing need to know accurate travel times for minor urban roads. Such information is crucial for tackling air quality problems, accommodate the growing number of cars and provide accurate information for routing. This study aims to address the aforementioned challenges by introducing a methodology, namely Similar Model Searching (SMS), to estimate travel times by using historical sparse travel time data. The SMS learns the temporal and spatial relationship between the travel time of adjacent links and utilise labelled data of similar models in order to improve its overall performance. The effectiveness of the proposed method is evaluated on a section of Leicestershire traffic network in the UK. The obtained results show that SMS efficiently estimates travel time of target links using models of adjacent traffic links.

No abstract