Back to Visual-tm home Back to overview Visual-tm models the behaviour of vehicles at a junction in a much better way than conventional methods. This offers the potential for more accurate modelling of congestion and junction delay. It provides for a much finer level of accuracy, much more detailed results and greater control over the model itself.
To calculate junction queuing and delay, Visual-tm uses the time dependent queuing theory and the junction capacity equations given in the Design Manual for Roads and Bridges (DMRB) which is the same method given in Arcady, Picady and Oscady. Working with TRL, we have also connected-in their junction design software so you can browse or re-design your junctions using Arcady, Picady and Oscady and re-test the effect of the new junction design on your network with Visual-tm.
To do this we have made a major innovation in the way assignment works.
Clock time Assignment
The day can be divided into consecutive time periods and the assignment model is run for each time period. Time periods are generally for one to two hours and represents the time period covered by the trips in the assignment trip matrix. Time periods are divided into ten-minute time intervals and junction and interchange turning movement flows are accumulated for each time interval and passed to the junction and interchange simulators which pass back turning delays in each time interval, for the next assignment iteration.Time intervals are divided into one-minute time slices. The trip matrix, which represents one time period’s worth of traffic, is divided into one time slice’s worth of traffic within the assignment process. Time slices are accumulated into time intervals for the highway junction simulation. Time slices are also accumulated and output to the too-late trip matrix for assignment in the next time period.
In order to provide the turning traffic volumes by time interval the assignment model keeps track of the clock time when assigned traffic volumes passed through each junction. This is called clock time assignment. It provides a much more rigorous way of modelling congestion because trip matrices are not considered to be all one lump of traffic which appears on all links for the whole modelled period. Clock time assignment divides the trip matrix into (typically one minute) time slices and assigns each time slice to the network.
Visual-tm’s assignment accumulates the junction turning flows by vehicle type and time interval and passes them to the junction simulation which calculates the queuing and delay.
Junction Capacity
The junction capacity for a particular traffic stream is generally a function of the traffic streams on other arms of the junction which may conflict with it – traffic streams which may vary over the course of the time period being modelled. This needs to be taken into account when calculating the junction capacity. Visual-tm’s junction simulation does this by taking the turning flows by ten-minute time interval supplied to it from the assignment model. The capacity for each stream of traffic is calculated from the empirical relationships used to design junctions which is embodied in TRL’s junction design software Arcady, Picady and Oscady.Having calculated the capacity of the traffic stream, Visual-tm can calculate the amount of traffic which will pass through the junction during one particular time interval. That traffic which does not get through the junction adds to the queue on the approach road.
Junction Delay
Visual-tm takes the queue at the beginning of the time interval, adds the effect of the traffic which does not get through the junction in the time interval and calculates the queue at the end of the interval using time dependent queuing theory. Queues from the end of one time interval are used as the starting queues at the beginning of the next time interval. Each time interval in the time period is examined in turn and the queues and delays calculated and output. The time dependent queuing and delay is also calculated from the empirical relationships used to design junctions in Arcady, Picady and Oscady. Visual-tm uses the same method – you can’t get a better method than that!What's Wrong With The Conventional Approach?
One common way of modelling junction delay relies on a user-defined relationship between the level of flow on the road approaching the junction and the delay incurred in passing through the junction. This method of modelling junction delay does not take into account the time dependent nature of traffic queue build-up and decline due to the traffic overcapacity history of a junction. This is the main source of delay and omitting it introduces major errors. The period where a junction flow is greater than capacity contributes to queue build-up so vehicles coming later will experience an amount of delay which depends upon the queue already build-up from earlier periods when its flow was greater than capacity. In addition, junction capacity depends upon the conflicts introduced by one traffic stream turning in front of another. This commonly-used approach can be done in Visual-tm if the user really wants to, but it is not recommended for congested networks.Some Other Advantages of Visual-tm’s Approach
Visual-tm’s is a detailed approach to modelling junctions which relies on current methodologies for modelling the time-dependent nature of queuing and delay and current methodologies for junction capacity and vehicle behaviour. It is completely consistent with junction design methodologies so when people come to design their junctions they don't have to go back and iterate with their assignment model to check whether the new junction is doing what it was supposed to do. Detail can be added-in where and when the additional detail is needed and the user can rely on existing methodologies for link-based assignment for the rest of the network. This approach to junction modelling is more detailed than most other assignment methodologies and therefore is likely to be more accurate.Modelling Long Trips
It is better at representing long trips and large study areas because of the effect of trips which start and/or end outside the modelled period. Trips are only added to accumulate the link flow volume for that part of their path that lies within the time period being modelled. Time periods can be concatenated so you can model the 6.00 to 10.00 am period in four 1-hour periods and carry-over queues, delays, and trips which haven't arrived at their destination yet. For example a large study area where the modelled time period is say from 8.00 to 9.00, some long trips could start well before 8.00 and arrive well after 9.00. A conventional assignment would have the whole trip assigned to the network, contributing to assigned links flows when in fact they would not be there between 8.00 and 9.00 – they would be there after (or before) the 8.00 to 9.00 time period. In London for example many drivers commute long distance into central London (or to Heathrow) starting well before the peak. This gives rise to a wave of traffic which starts early (eg 6.00 to 7.00am) outside M25, reaches outer London a little later (eg between 7.00 and 8.00) and inner London during the peak proper (eg between 8.00 to 9.00). This effect can now be modelled with Visual Transport Modeller version 3 using the clock time assignment methodology and the junction congestion and delay simulation.Modelling Congestion Charging and Air Quality
With Visual-tm, you can now model congestion charging properly and apply a charge based on the amount of time a vehicle is queuing. In Visual-tm, delay is the amount of time spent queuing at a junction and delay time has its own generalised cost parameter which is used for applying the congestion charge as a monetary value per delayed minute. Furthermore the network can be skimmed to get a matrix of delay time for each origin-destination pair so you can see exactly what each origin-destination movement pays what amount. This skim can be put into the mode choice and distribution models so as to carry the effect of the congestion charge through to changes in mode and destination. Air quality depends upon the time vehicles spend travelling and queuing which are output in the skims so Visual-tm’s outputs provide better data upon which to forecast air quality.
Back to Visual-tm home Back to overview