This paper proposes a short-term rolling horizon framework for the within-day prediction of travel times on links with and without point detectors (referred to as observed and unobserved links respectively) along a selected path covered in the Hong Kong journey time indication system (JTIS). In Hong Kong JTIS, the number of point detectors on major roads is usually limited due to the financial budget and site constraints in the densely populated urban area. However, the prediction of the travel times on urban road corridors particularly on the links without point detectors is also valuable to road users and traffic authorities. This paper proposes a 2-stage framework based on functional principal component analysis and maximum likelihood estimation method to predict the mean and standard deviation of the travel times on the study path and observed links as well as unobserved links once every 2 min for the next 30 min. An urban road network in Hong Kong is selected as a case study. The prediction results are validated using an independent dataset from JTIS, demonstrating the practical applicability of the proposed framework.
- Functional principal component analysis
- Maximum likelihood estimation
- Multi-source traffic data
- Travel time prediction
ASJC Scopus subject areas