PRINT BOOKMARK

Active travel counting conundrums

James Datson | 12 Oct 2015 | Comments

For decades local authorities have invested heavily in collecting motor vehicle count data on their networks. Such data is the life blood of our transport models and is vital to help traffic managers make decisions on how they manage the highway. But to date many authorities have struggled to make the case for monitoring the behaviours of an often equally important user – the cyclist and pedestrian even though understanding these users could help attract significant funding.

Why is this? Well the obvious answer is that these active travel modes make up a relatively small proportion of highway users and therefore don’t warrant data collection investment. Other arguments are linked with the challenges of obtaining statistically significant results on what can be a relatively small number of journeys. Furthermore, there can be many measurement points that need to be monitored to get a good understanding of the route choices made by pedestrians and cyclists and this can make monitoring too expensive. These types of arguments can be bundled up to often mean that for reasons of “proportionality”, active mode data collection doesn’t happen. But the counter argument is that to make more liveable towns and cities we need to support active travel modes, and political will to do this has been sustained for some time, from the Prime Minister down.

From DfT and Highways England, down to individual borough councils, we consistently see policy aspiration to do more for active travel modes. So is now the time to fill the gaps in the data used to understand our pedestrians and cyclists? We think so, but key questions must be asked around which technologies work and does the business case stack up to use them.

One progressive authority in this area is Transport for Greater Manchester where we have recently won a commission to trial a new type of vision based count technology. The technology has the potential to count pedestrians and cyclists in near real-time by recognising the shapes of these users just like a human enumerator would. The technology uses machine learning to learn what a pedestrian and cyclist looks like amongst a mix of traffic and because the marginal costs are low we may see this type of technology being used to offer a Data as a Service model to our clients. Even if a DaaS model doesn’t gain traction, the technology could offer advantages over traditional data collection in terms of reduced error rates, larger sample sizes and more robust data to support to our models.

Ultimately such technology may help our clients develop a better understanding of the preferences and behaviours of cyclists and pedestrians. Better information will undoubtedly result in more informed decisions and help the active modes be better represented our investment decisions.

We are looking to build the capabilities of this type of technology through a number of future trials and in parallel with this investigate the business case for investing in DaaS type services. What is particularly interesting is the potential to find new ways of funding more sophisticated data collection e.g. by using existing CCTV assets. And we may find new revenue streams for authorities to sell their data to third parties such as retailers and those responsible for monitoring the security of our urban areas.

How do you see better data collection technology supporting the needs of pedestrians and cyclists in our cities?

> To continue the discussion on Intelligent Mobility, please join our dedicated LinkedIn Group