A computationally efficient two-stage method for short-term traffic prediction on urban roads

Abstract

Short-term traffic prediction plays an important role in intelligent transport systems. This paper presents a novel two-stage prediction structure using the technique of Singular Spectrum Analysis (SSA) as a data smoothing stage to improve the prediction accuracy. Moreover, a novel prediction method named Grey System Model (GM) is introduced to reduce the dependency on method training and parameter optimisation. To demonstrate the effects of these improvements, this paper compares the prediction accuracies of SSA and non-SSA model structures using both a GM and a more conventional Seasonal Auto-Regressive Integrated Moving Average (SARIMA) prediction model. These methods were calibrated and evaluated using traffic flow data from a corridor in Central London under both normal and incident traffic conditions. The prediction accuracy comparisons show that the SSA method as a data smoothing step before the application of machine learning or statistical prediction methods can improve the final traffic prediction accuracy. In addition, the results indicate that the relatively novel GM method outperforms SARIMA under both normal and incident traffic conditions on urban roads.     

Analysis of intended bus usage

In order to plan bus operations, it is necessary for transit planners to understand what factors may influence travelers’ choice of buses for travels within a city. The proposed method involves various scenarios of a hypothetical bus operation which was rated by a group of individuals.

Analysis of Covariance technique is employed to analyze people's sensitivities to their perceived levels of bus service characteristics. The technique involves:

1. testing for the significant effects of varying levels of service characteristics upon people's intentions to use bus service, and

2. assessing differences among various population segments in their sensitivity patterns towards bus service characteristics.

Results from the application of the technique to attitudinal data collected by the Orange County Transit District indicate that bus service characteristics do influence, independently and jointly, respondents’ stated intentions to use buses.

Sensitivity pattern differed across the five homogeneous segments identified in an earlier research based on socioeconomic characteristics.

One segment (an older, predominatly male population segment with higher home ownership level and lower income than the rest of the sample) was relatively insensitive to changes in bus fare and was influenced by changes in headway independent of changes in access distance. Another segment consisting of fewer registered voters with lower education also exhibited similar independent impact of headway and access distance.

The technique is especially useful in reducing a large number of proposed alternative bus systems to a smaller set for further planning considerations by specifying the ranges within which variation of service characteristic would cause substantial changes in the intended usage responses.     

The influence of alternative data smoothing prediction techniques on the performance of a two-stage short-term urban travel time prediction framework

This article investigates the impact of alternative data smoothing and traffic prediction methods on the accuracy of the performance of a two-stage short-term urban travel time prediction framework. Using this framework, we test the influence of the combination of two different data smoothing and four different prediction methods using travel time data from two substantially different urban traffic environments and under both normal and abnormal conditions. This constitutes the most comprehensive empirical evaluation of the joint influence of smoothing and predictor choice to date. The results indicate that the use of data smoothing improves prediction accuracy regardless of the prediction method used and that this is true in different traffic environments and during both normal and abnormal (incident) conditions. Moreover, the use of data smoothing in general has a much greater influence on prediction performance than the choice of specific prediction method, and this is independent of the specific smoothing method used. In normal traffic conditions, the different prediction methods produce broadly similar results but under abnormal conditions, lazy learning methods emerge as superior.     

A methodology for planning bus service

A methodology to assist transportation planners in designing bus services is developed. The methodology is most relevant for use in locations where bus service of the type being studied does not currently exist and therefore no information is available on past choice behavior, or in instances when transferability of travel models estimated in another location is difficult. The methodology assesses the sensitivity of bus service characteristics upon intended bus usage using survey data collected in Orange County, California, by the Orange County Transit District (OCTD). The methodology is based on a nonparametric statistical test developed by Kolmogorov and Smirnov.Scenarios describing hypothetical operations of bus service are presented to survey respondents who indicate their intended levels of bus usage under each situation. Significant differences between the response distributions associated with pairs of scenarios are identified and potential ridership levels, as bus operations become more favorable, are assessed. Various user segments are then identified on the basis of their levels of intended bus usage and the corresponding marketing implications associated with each segment are discussed.