Private vehicle network

The private vehicle network was the simplest of all the transportation networks to create. Starting with a GIS layer of road centreline locations, the appropriate attributes were set in order to model the travel of private vehicles through the road network.

Firstly a field was added to hold the length of each edge (road segment) in the network. Then this field was populated using the GIS to measure the length of each network edge.

Each edge in the network was then attributed with a ‘best estimate’ of the average travelling speed based on the road hierarchy. The average speeds selected for the model of the Christchurch private vehicle network are shown in table 14.1.

Table 14.1    Modelled speeds for the Christchurch private vehicle network

Road type

Modelled speed (km/hr)

Private road

20

Local road

30

Collector

30

Minor arterial

40

Major arterial

40

Motorway

70

 

The average speeds applied to each edge in the network should be based on actual or modelled values when that information is available. For example in the UK the accessibility indicators use the same data sets that are used in dynamic driver information systems. From GPS systems in cars and the speed of travel of mobile phone networks along each road link there are now comprehensive data sets available showing the speed of travel of vehicles on most road links. Only very lightly trafficked roads do not currently have associated speed measurements. GIS systems like Google and Bing Maps now show live traffic speeds and the data can also be purchased for accessibility modelling. In the UK, DfT has purchased data for the calculation of accessibility indicators since 2009. The speeds at which vehicles travel through networks can also be established from survey data using number plate recognition systems and this can be used to calibrate average speeds for different categories of vehicle.

Other useful information to assist with allocating correct speeds to each road link includes information such as delay at intersections and link speeds, which can often be extracted from traditional traffic assignment models. When the estimation is based on a categorised system, such as the road hierarchy or speed limit, a reduction factor should be applied to account for the effects of delays generated by intersections, congestion and traffic management devices. Further application of this technique could include areas such as slow or congestion zones where slower network speeds are evident such as in a city centre.

Finally a field was added to the network dataset to hold the calculated traversal time for each edge. This was measured in seconds and was calculated from the length of the edge in metres and the traversal speed in km/h as follows:

Edge traversal time (s) = edge length (m)/edge traversal speed (m/s)

Where it was necessary to model one-way streets, an attribute was added to the dataset to record if the edge was traversable in the direction it was digitised, opposite to the direction of digitising or in both directions. This attribute enabled modelling of one-way restrictions on travel across the network.