Updating Traffic Estimates - A review of industry practice
by Mike Tapper - for RIMS
RIMS (Roading Infrastructure Management Support) has recently completed a project to review industry practice for updating traffic count estimation. Currently there is no guideline for undertaking traffic count estimates. The purpose of this report was to research current methods used in the industry for updating traffic estimates in RAMM. Part of the report also commented on these various methods to form a knowledge base to assess a best practice guide. It is expected these guidelines will be released later in 2020. This programme is also running in parallel with a REG project to upgrade the traffic count estimation (TCE) tool in RAMM.
It is important to note that these guidelines are targeted at local authority networks. State Highway networks require a completely different approach.
The industry review covered current practice undertaken by Beca, Stantec and WSP. It also covered methodology undertaken at Auckland Transport internally. The review outlined the current process within the RAMM TCE tool plus proposed improvements to the tool outlined in the project led by REG to update the TCE tool.
The findings showed that methodologies varied quite significantly across the different parties. This depended on both the nature of the networks and each proponent’s different experiences and preferences. There was a similar theme through some methodologies in terms of overall framework though, even though different approaches were adopted at stages through that process.
The recommendations of the report involved tying together those findings into a possible proposed framework from which three findings were identified
- Establishing the context for traffic monitoring updates
- Desirable qualities for an estimation update process
- Recommended Estimation Update Process Framework
For context, understanding how confidence in our estimates should and does vary is important. Confidence should vary with the importance of our road sections on our network. Important road sections should be counted more frequently, both in terms of time between counts and frequency along the road corridor. This is because we are seeking more confidence in the accuracy of these estimates. Conversely lower ranked road corridors and sections will be counted less frequently. Hence we will have less confidence in the estimates resulting from these sections, as we require less confidence. A difference in estimate of 20 vpd to 50 vpd on a low volume road is much different to accepting a range of 10,000 vpd to 15,000 vpd on a high volume road section for example.
This then touches upon the second aspect of context. Our estimates need to provide sufficient confidence to know the journeys travelled across our network i.e. VKT. But we also need sufficient confidence in our estimates for each road section across our network. Thirdly, we need to understand vehicle mix across our network also. Therefore our estimation process framework (and count strategy) need to bear in mind these objectives also.
Attributes to look for in our estimation process are:
- Provides satisfactory confidence in the accuracy of the network estimates
- Easy to set up
- Simple to maintain
- Simple to undertake
- Manual tasks minimised
- Covers volume and mix
- Aligns with the traffic count strategy best practice approach (core and rotational sites)
- Scalable i.e. can be used as a base from which more customised processes can be introduced
The framework process is as recommended is as detailed below. Various approaches have been used for each step and these are discussed in the report:
New Sections/Sections with no estimate: Before starting the process, checks should be made that each carriageway section has an estimate.
Using Links: Links reduce the number of similar performing sections, improving and focussing the sampling analysis for completing a traffic count site strategy and programme.
Using Groups: Traffic groups are collections of links (or sections) that have similar profiles. They are used to inform sections without relevant volume of mix information with data from those sections within the group with sufficient relevant count data.
Estimating Sections with Counts since the last estimate: The best data to update an estimate is from a valid count on that section of road since the last estimate.
Estimating links with adjacent counts: Next best is to update the estimate based on counts adjacent to the section in question.
Estimating links with no counts: If no relevant count history is available, estimate updates in volume can be determined by a number of different approaches, often utilising the trends and changes within the grouping of similar road sections or links.
Traffic Mix: A similar approach in followed for updating traffic mix estimates i.e. directly updating links associated on or near count sites and then adjustments by averaging mix over the remaining links in the group.
The research undertaken in this report will be used to document a guideline to best practice for updating traffic estimates. There is also an update underway to the RAM TCE tool which will reflect aspects of the guidelines also.
Thanks must go to the contributors to the review who cooperated freely in sharing their approaches to traffic estimation to assist in improving this key aspect of asset management.