Monash University

Restricted Access

Reason: Access restricted by the author. A copy can be requested for private research and study by contacting your institution's library service. This copy cannot be republished

Combination Effects of Public Transport Priority Measures

posted on 2017-02-13, 01:16 authored by Long Tien Truong
Public Transport Priority (PTP) is crucial to improving Public Transport (PT) travel time and travel time reliability. PTP can be categorised as Road-Space Priority (RSP), e.g. Queue Jump Lane (QJL) and Dedicated Bus Lane (DBL), and Transit Signal Priority (TSP). To increase its effectiveness, TSP may be combined with RSP, e.g. TSP combined with DBL (TSPwDBL) or TSP combined with QJL (TSPwQJL). There has been a growing body of literature on the design and operation of PTP measures. Yet, no research has been undertaken to establish whether combining TSP with RSP measures creates an over-additive effect where the combined treatment effect is greater than the sum of their separate effects. In addition, it remains unclear whether providing PTP measures at multiple locations, e.g. road sections and intersections, along a signalised arterial creates a multiplier effect where the effect from providing PTP measures at multiple locations is greater than the sum of effects from providing PTP measures at each of those individual locations.
   This thesis therefore aims to provide an in-depth understanding of the combination effects of PTP measures. A research approach with six components is adopted. The first four components focus on methodological developments, which are then applied for understanding the combination effects in the last two components.
   Traffic Micro-simulation Models (TMMs) are selected as an experimental tool for this research. The first component develops new methods to calculate the Minimum Number of Runs (MNR) required to estimate multiple Measures of Performance (MOPs) at an overall confidence level, using multiple comparisons in statistical theory. Numerical analyses based on an example corridor suggest that new methods provide an improved means of assessing the statistical accuracy of multiple MOPs.
   Given the impacts of signal coordination on the performance of PTP measures on arterials, it is crucial to investigate their combination effects when signal offsets are optimised. Hence, the second component proposes new offset optimisation models, which consider randomness in bus dwell time and PTP measures such as DBLs and QJLs. In these models, bus and general traffic delays are estimated using kinematic wave theory. Validation using TMMs demonstrates the effectiveness of these new models.
   Bus delay functions can be used as a PTP planning tool and to examine the effects of combining TSP with RSP measures at an intersection level. Thus, the third component proposes a set of kinematic wave theory-based delay functions to estimate bus delay effects for a range of PTP measures (i.e. DBL, QJL, TSP, TSPwDBL, and TSPwQJL). Delay associated with bus acceleration is also approximated. Validation using TMMs indicates an acceptable level of agreement between delays obtained from new delay functions and TMMs.
   While this thesis focuses on the effects of combining conventional (local) TSP control at multiple intersections, it is also prudent to examine whether a coordinated TSP control model, which considers bus progression through multiple intersections, performs better than providing the conventional TSP control model at those intersections. The fourth component therefore develops a coordinated TSP control model considering arterial progression and stochastic bus arrival time. Extensive TMM experiments show that the coordinated TSP control generates an additional reduction in bus delay of 11% to 19% when compared to conventional TSP control.
   In the fifth component, kinematic wave theory-based analyses, together with new bus delay functions, suggest that at an intersection level, the effects of combining TSP with RSP measures on bus delay savings are smaller than additive if there is no nearside bus stop and the traffic condition in the base case is under-saturated. With a near-side bus stop, the combined treatment effect on bus delay savings at an intersection level can be over-additive, depending on various factors, e.g. dwell time and distance from the bus stop to the stop line. In addition, analytical results indicate that at an arterial level, the combined treatment effects on bus delay savings can be over-additive with suitable signal offsets. These results are confirmed by a TMM case study. The over-additive effects of TSP with DBLs or QJLs result in an 8-28% increase in bus delay savings, when compared to an additive effect.
   In the sixth component, TMM results reveal that providing PTP measures (i.e. QJL, TSP, TSPwDBL, and TSPwQJL) at multiple locations can create a multiplier effect on one-directional bus delay savings, particularly when signal offsets provide bus progression for that direction. In general, the multiplier effect results in a 4-8% increase in bus delay savings for each additional road section or intersection with PTP measures, when compared to a Constant Returns to Scale (CRS) effect. A reason for the multiplier effect on bus delay savings is that PTP measures reduce the variation in bus travel times. This enables signal offsets, which account for bus progression, to perform even better, particularly when the variation in bus dwell times is not high.
   This thesis provides six original contributions to knowledge in line with the six research components. Overall, it is concluded that combining PTP measures, either TSP with RSP measures or PTP measures at multiple locations on signalised arterials, will help to improve travel times and reliability for PT passengers, ultimately enhancing the PT user experience.


Campus location


Principal supervisor

Graham Currie

Additional supervisor 1

Chris De Gruyter

Additional supervisor 2

Kun An

Year of Award


Department, School or Centre

Civil Engineering


Doctor of Philosophy

Degree Type



Faculty of Engineering

Usage metrics

    Faculty of Engineering Theses