Blurred highway

Does slower driving make us go faster?

21 December 2018
From our ‘Thinking outside the box’ series

Traffic management measures that try to prevent traffic flows on motorways from breaking down typically reduce vehicle speed. Michiel Bliemer & Mark Raadsen discuss, are we in fact faster by driving slower?

Many drivers use motorways (tolled or untolled) in order to have a fast trip without traffic lights. Often these motorways are congested with slow moving traffic. There exist however traffic management measures that try to prevent traffic flows on motorways from breaking down (avoiding congestion as a result). However, these measures typically reduce vehicle speed and begs the question: are we in fact faster by driving slower?

As an example, consider ramp metering. Ramp metering is a widely used traffic management measure in which access to the motorway is actively restricted to prevent congestion. The idea being that if too many vehicles enter a motorway simultaneously this leads to traffic breakdown, simply because the merging of these new vehicles with motorway traffic causes friction in the form of lane changes and sudden braking. Consequently, all vehicles on the motorway suffer congestion delays if this friction becomes too severe. Ramp meters are traffic lights near the on-ramp of the motorway that let only one or two vehicles access the motorway at the time. Detectors on the motorway continuously count the number of vehicles and are used to optimise traffic flow by dynamically adjusting the ramp meters. By letting some vehicles wait in a queue at the on-ramp, all traffic on the motorway is kept in free-flow as shown in this animation, which – in the end - benefits all drivers, even the ones that are temporarily forced to wait.

However, even with ramp metering in place there still exists the possibility of traffic breaking down on motorways. If the motorway is near capacity (i.e., when counting around 35 passing vehicles per minute per lane) then vehicles drive at high speeds with only small distances between them, making traffic flow very unstable. A driver that suddenly forcefully brakes for a slower vehicle in front (possibly merging from another lane) can set off a cascading effect called a wave in which following vehicles also have to brake, resulting in traffic flow break down. Many people will have experienced such stop-and-go waves with so-called phantom jams in which there is suddenly congestion for no apparent reason. These phantom jams were already described in 1972 by Professor Wilhelm Leutzbach from the University of Karlsruhe, and a nice video demonstrates these phantom jams.

Professor Katsuhiro Nishinari from the University of Tokyo has suggested that drivers on a motorway keep a constant and sufficiently large distance to the lead vehicle when driving in order to prevent traffic flow breakdowns. The idea behind this advice is that it creates buffer space in order to prevent sudden braking from cascading. While keeping more distance will indeed help in reducing phantom jams, it unfortunately also decreases throughput. For example, keeping a headway of 40 meters or more when driving 60 km/h, as suggested by Professor Nishinari, results in a throughput of only less than 22 vehicles per minute per lane, much lower than the road capacity of 35 vehicles per minute per lane. While this will keep traffic on the motorway flowing, it means that less vehicles can be admitted to the motorway by the ramp meters. This will result in long queues on on-ramps that will likely spillback onto other intersections and cause problems on urban roads. In other words, increased traffic flow stability can lead to decreased traffic flow efficiency.

Improvements in both traffic flow efficiency and stability can be achieved by reducing the maximum speed on motorways during peak periods to 80 or 90 km/h. At this speed, traffic flow is maximum at 35 vehicles per minute per lane while it will also decreases speed differences between vehicles. This creates more homogeneous traffic flow in which sudden braking by drivers is significantly reduced.

Further, reduced speed limits also reduce speed differences between lanes such that drivers have less incentive to change lanes, which again increases traffic flow stability. Daganzo’s behavioural theory of multi-lane traffic flow distinguishes two types of drivers, namely slugs that prefer to be in the slower lane(s), and rabbits that prefer to be in the faster lane(s). Because of differences in preferred driving speed, some lanes are not used efficiently (except when there is an optimal mix of slugs and rabbits). Reducing the speed limit forces rabbits to adopt the same speed as the slugs, removing the urge to overtake and resulting in traffic flow being spread more evenly across lanes. The end result of this measure is an increased efficiency of the motorway as a whole. In another think piece the same theory was used to argue that forcing everyone to stand (not walk) on an escalator can improve throughput and reduce delays.

To summarise, we can all be faster by slowing down during peak periods. This counterintuitive result stems from the fact that motorway capacity is used most efficiently when traffic flows and speeds do not vary much over time, which can be achieved through ramp metering and reducing speed limits. Given that traffic flow is composed of individual drivers, it is the duty of each driver to ensure that they cause minimal disruption to traffic flow, since small disruptions can have large consequences for everyone.