On first consideration, road signs seem critical to the proper functioning of the road system. Another point of view is that like ‘paying with cash’ they are a relic, no longer required in the world of connectivity, satnavs, driver assistance systems, and (soon) fully automated cars which drive themselves.
Of course road users require information. The question is, do we still need physical road signs to provide this? In this short essay, I’d like to try and answer that question, and along the way consider road signage from a number of perspectives.
Roads signs, surprisingly, are a rite of passage
We probably don’t think much about road signs until we learn to drive. When we start learning though, they take on great importance. They become one of the barriers to the rite of passage that is obtaining a driving licence. To pass the driving theory test, a learner driver will need to prove that they understand the meaning of (a sample of) road signs. Until this barrier is passed, even booking a practical test is not possible, let alone passing one. The reasoning behind this, one might think, is that the driving test is measuring critical skills which people need in their later driving careers. Well, not necessarily. Shinar et al. (2003) studied sign comprehension in multiple countries and groups including novices and older, more experienced drivers. They found that more experienced drivers almost always performed worse than novices; the levels of sign comprehension when learning would appear to be more than is needed in later driving.
People actually use road signs (sometimes)
What about warning drivers of hazards on the road ahead? Here there does seem to be a plausible role for signs to play. For example, some clips in the UK’s hazard perception test make use of signs which, when combined with a vehicle in the scene, can be used to predict a developing hazard. We know that drivers use signs to guide their driving style in risky situations too. For example in one of my earliest projects at TRL ( Helman et al, 2010) we had drivers in an instrumented vehicle drive a part of the A377 in Devon, through a number of bends known to have a high risk of vehicles leaving the road. The amount of warning signage and other treatments on a bend led to drivers choosing slower speeds, even when the effect of the actual geometry of the bend was taken into account; in other words, people slowed down when signage warned them to. Neat.
The importance of road users being able to see road signs is not lost on cutting edge projects looking at vehicle automation either. For example in the Helm UK Truck Platooning project, TRL is assessing what impact platoons of lorries will have on the ability of other road users (and the truck drivers themselves) to see road signs and, critically, whether that matters.
Automation on vehicles will not eliminate the need for signs soon
One line of thought is that we will no longer need road signs at all when vehicles begin to use greater levels of automation than in (relatively simple) platooning. However in the forthcoming EU General Safety Regulation the Intelligent Speed Assistance (ISA) systems included will actually require visible speed limit signs to be readable, even though manufacturers can use other methods such as databases of speed limits to help (see Seidl et al., 2017 for the report TRL prepared to underpin this regulation).
Things get interesting when we look at forthcoming ‘truly automated’ vehicles which drive themselves. Some companies, notably TESLA, are working on camera-based systems while others are focusing on LIDAR. The mix of technologies used in sensor suites on these vehicles may have an impact on what kinds of signage we need on future roads (camera-based systems, for obvious reasons, work best with current sign designs). TRL has been involved in a number of recent projects looking into this. For example in the MOVE_UK project, we looked at an automated sign recognition system on the test vehicle and found that for around 1 in 14 speed limit signs, real-world detection rate was between 50% and 80% - hardly sufficient to support ISA. The project uncovered a range of reasons for this including obscuration, and poor sign orientation (for example side-road speed limit signs being visible from a main road). This is interesting; we require signs to be usable by technologies which cannot seem to read them as effectively as the drivers the systems are replacing. In journeys forward, there are usually some steps back.
Technology innovation is quick to start, but slow to complete
Trials with automated vehicles are all very well but they are not the only thing we should be focused on. Trials to get rid of sign clutter through use of technology are also to be welcomed; Highways England for example has been involved in trials to see if information that would normally require physical signage can be directly ported to screens inside vehicles. The A2/M2 Connected Corridor and ‘UK-CITE’ projects are both examples. If we can get rid of physical signage this also has benefits for road safety – people who lose control of their vehicles will have fewer objects to collide with at the roadside if we can minimise unnecessary signage.
There is a problem though. New technologies can take hold quickly, but complete market penetration takes ages. An aside will illustrate this point well. This week I received a document through my letterbox called a ‘phone book’; the analogy here is clear, and even though the publication’s physical ability to pass through my letterbox is an indication of its decline in usefulness as people migrate to online information, it is still here. I expect it be the same on the roads; some people will not have access to vehicles that can benefit from ‘digital sign technologies’ and it will take a whole refreshment of the vehicle parc to make such systems ubiquitous.
Do we still need road signs? Some of us do, at least some of the time. Even the initial generation of connected and digital-ready vehicles will to some extent. And after everything is ready for 100% digitisation, someone somewhere will still want to look at a road sign, to navigate to a restaurant, quite possibly one that they looked up in a phone book, and at which they will pay for a meal with cash.
References:
Helman S, Kennedy J and Gallagher A (2010). Bend treatments on the A377 between Cowley and Bishops Tawton: final report (PPR494). Crowthorne: Transport Research Laboratory.
Seidl M, Hynd D, McCarthy M, Martin P, Hunt R, Mohan S, Krishnamurthy V and O’Connell S (2017). In depth cost-effectiveness analysis of the identified measures and features regarding the way forward for EU vehicle safety. Brussels: European Commission. DOI: 10.2873/748910
Shinar D, Dewar RE, Summala H and Zakowska L (2003). Traffic sign symbol comprehension: a cross-cultural study. Ergonomics, 46(15), 1549-1565.
