Monday 13 September 2010

Junction design for safer cycling

Another of Mark Wagenbuur's excellent videos, showing the design of a typical traffic light junction in the Netherlands and how it differs from the design in other countries.

This type of design not only increases safety of cyclists, but also the important subjective safety required to make people want to cycle, and want to let their loved ones cycle.

The Netherlands has the highest cycling rate in the world, so more vulnerable road users than other countries, yet also the safest roads in the world. It is infrastructure like this which has achieved such a level of safety for cyclists.

Note that in most cases a cyclist can make a right turn without stopping at the traffic lights. This is usually the case even where there isn't such an obvious cycle path on which to do it. As a result, even junctions which require a cyclist to stop twice to make a left turn provide cyclists with equal average speed through a junction to when riding on the road. At many modern junctions it is not necessary to wait twice to turn left, so cyclists get to stop less frequently than drivers.

This blog has more examples of traffic light junctions and cycle paths in the Netherlands (and elsewhere).


Nick said...

This is a great video, for sure, that demonstrates many of the benefits of this design. I noticed, however, one discrepancy between Mark's narrative and the video example:

At 2'41'', Mark says that the cycle path has its own traffic light that "has a different green cycle from right-turning cars," but the clip with the gentleman at 3'08'' shows him driving through the intersection at a green at the same time as a car tries to turn right. It is clear in the video that both signals change to green at the same time.

The situation is without a doubt an improvement over what we see here in Canada, but a conflict still exists between cyclists and right-turning cars, especially as both approach a junction where the light is already green.

David Hembrow said...

Nick, I was surprised to see that too. We don't have any junctions like it around here. My guess is that it's an older junction yet to be upgraded to modern standards.

However, it's not really a bad junction. There is considerable separation between the cyclists and motorists at this point, cyclists do not start with drives, but are well ahead and closer to the junction itself when the lights go green. That increases safety. An accidental collision would be difficult to arrange. A driver would have to take that tight right turn at speed and deliberately drive at cyclists going across his path.

A lack of separation of cyclists going straight on from drivers turning right is something I've criticised before about the Danish approach which seems to maximise conflict at junctions.

Normally cyclists are more completely separated than this - often by a separate cycle of the lights for cyclists. I have several examples on the blog of other junctions.

Jon Bendtsen said...

It is only some Danish lights that do not separate cyclists from drivers. You even featured one of those that do separate cyclists and drivers. The one with those blinking lights in the ground you wondered about.

David Hembrow said...

Jon, if that junction properly separated cyclists and drivers it wouldn't need flashing lights in the ground to warn drivers. That was in fact the basis of my complaint about it. It's dressed up as an advance and something for the rest of the world to copy when actually it's an example of poor design.

Mark W. said...

Yes it is the exceptions that confirm the rule, as the Dutch say...

In that one example indeed the cyclists gets green only a fraction of a second earlier than the car. But what happens next is very telling so I kept it in.

The cyclist doesn't notice it is green right away so starts riding very late. Yet he is still passed the junction much earlier than the turning car that has a long way to go before it reaches him. The car can see him all that time and keeps the speed down in anticipation. So even in this particular exceptional case where they get green at the same time, the cyclist is still better off than he would be on traditional junctions.

As a whole: everything I state is 'in general', 'usually' etc. Because no two junctions are the same. There is also still a legacy of old junction design from the 1960s and 1970s so exceptions to the ideal solution do still exist.

christhebull said...

Dutch towns do, however, have junctions with a mixture of separate tracks and filter lanes (wide filter lanes though, not the crappy UK style ones by the kerb that encourage stupid filtering...), but they are mostly where a minor road which may not have cycle lanes has a junction with a major road with separate tracks.

A question - on roads without cycle paths, do cyclists typically launch themselves off the kerb when turning left (mimicking the manoeuvre for larger junctions with cycle paths), or move towards the centre first? Or would that typically depend on the location and skills of the individual rider?

Nick said...

"Yes it is the exceptions that confirm the rule, as the Dutch say..."

Of course it's the exception I notice right away only because the consistent quality of Dutch infrastructure (as I have seen through blogs like this one) makes me weep with envy.

Problems like the one I highlight in <a href=">this video</a> are, sadly, the norm here in Ottawa.

I will be visiting Amsterdam in a few days so I'm looking forward to bringing back images of cycling infrastructure for the blog. Canada needs to learn how much better it can be.

Slow Factory said...

Two questions:

1 - In general, what happens in design conflicts between cycling and pedestrian infrastructure? For example, if a busy crossroads - where both streets have speed limits above 30km/h - needs stuff as shown in this video, but due to existing location of buildings the pedestrian space cannot be narrowed (what is the NL standard for that?) Does the "car part" of the street get narrowed?

2 - My favourite curiousity these days: What are the real objections in NL to voluntary use of the "car part" of the road by cyclists on roads/streets where separated paths are implemented, which makes it illegal to cycle in the "car part"? To clarify, I would prefer carfree streets but accept good separation like in NL, and do not think any cyclists should be forced into dangerous situations. I think that so-called "vehicular cycling" ignores reality BUT this, uh, separation philosophy can lead to the acceptance of motor vehicles, forever (and their emissions kill more than collisions).

Cycling infrastructure is not great here in Berlin, and it very clearly supports the unique - by German standards, at least - opportunity here to ride in the "car part" of streets even if there is a separated path. This is definitely no solution for people intimidated by non-separation, but it can provide increased capacity if, for example, trees cannot be cut down, a whole line of parking cannot be removed or a moving lane cannot be removed (the latter two of course only because of political reasons) and pedestrian space cannot be narrowed.

It is about freedom. If injuries go up in relation to higher cycling rates due to the wider application of this, it only means that motorized vehicles need to be further slowed and so on.

I accept that there are advantages to trams, buses and emergency vehicles not having conflicts with cyclists, but I don't think that this is an excuse for separation on all faster streets.

Anonymous said...

1 - In general, what happens in design conflicts between cycling and pedestrian infrastructure?

Pedestrian infrastructure always comes first. A road without decent pavements is unthinkable. If necessary, cyclists will have to share roadspace with motor traffic, or the road will be made one-way.

A question - on roads without cycle paths, do cyclists typically launch themselves off the kerb when turning left (mimicking the manoeuvre for larger junctions with cycle paths), or move towards the centre first?

You wouldn't launch yourself off the kerb, as you wouldn't be on the pavement to begin with. Most junctions without any facilities are between two 30 km/h or otherwise low-traffic roads, where you can generally cross at your leisure. I guess the place where this comes up most is at a junction of a major and a minor route. At such a junction, there would generally be traffic islands on the major road. This allows a cyclist coming from the major road to either take the lane and then pause between the two traffic islands, or to stop at the roadside and make the crossing in two stages. Similarly, a cyclist coming from the minor side street would be able to cross in two stages when making a left turn.

2 - My favourite curiousity these days: What are the real objections in NL to voluntary use of the "car part" of the road by cyclists on roads/streets where separated paths are implemented, which makes it illegal to cycle in the "car part"?

This is actually a serious issue, especially to recumbent and velomobile riders. Occasionally you will see non-mandatory cycle paths alongside roads, but many motorists are not aware of this and will get irritated at a cyclist sharing the road when this can be avoided. However, I do use the roadway when the cycle path is full of school-age kids or people loading their shopping into their panniers - even when it's not legal to do so - and I don't get much hassle from motorists. I think the main objection by motorists would be that they are liable if they hit me, even if I'm illegally using the main roadway. It is in both their interest and mine that we do not get into a collision, and me choosing to ride in a more dangerous position increases both our risks.

Slow Factory said...

Anonymous: Thanks for clarifications! In regards to the "voluntary vehicularism" I mention, it is indeed an issue of speed-separation as well as weight/mass/force separation.

So I am curious about the non-mandatory cycle paths you mention. Seems like one or the other needs to be the default (so there is only one set of signs and simpler education), or perhaps it can be decided on inside/outside city limits.

Or do "car parts" of roads have a minimum speed limit and can the regulations behind this be modified to include cyclists?

So, yes!, why should wonderful and efficient recumbent/velomobilers be impacted by a separation methodology based on a rather crude evaluation of typical velocity differentials between motorised and non-motorised conveyances?
Would a change in laws at least in certain areas support a greater uptake of these inherently faster machines?

Liability rules should remain as is. I would go further with the related philosophy and suggest that private unguided vehicles of all sorts should be required to travel more slowly than lighter ones, at least in cities. OK, that is really pushing things but I am the eternal patient optimist.

As I have mentioned in related discussions in this and other blogs, I grew up in a strictly vehicular environment and actually feel less safe than when I was younger moving that way, and I wonder how much this has to do with living in places with separated infra.

BUT I also have this feng shui thing about being the middle of the street whenever possible. The balanced road diet, if you will.

Frits B said...

In addition to Anonymous' information: Holland has 2 kinds of separate cycle paths and also 2 kinds of cycle lanes. Cycle paths are either mandatory (marked by a round blue sign with in white the outlines of a bicycle) or non-mandatory (marked by a rectangular sign, usually dark blue, with text "Fietspad" or "Rijwielpad"). There is a tendency to make all new cycle paths mandatory. The non-mandatory variety is usually found alongside roads with a speed limit of 70 km/h or less; all roads with higher speed limits have mandatory cycle paths - and if they don't, cyclists are banned from them anyway because the road is for motorized vehicles only, and signposted as such.

Cycle lanes can be marked by continuous or intermittent lines. In both cases the cyclist is supposed to stay in the cycle lane, but a continuous line implies an obligation for the cyclist to do so, and for cars to stay out of the cycle lane. Intermittent lines tell both parties that they can cross into each other's lanes if necessary. This is all theory, however; lines usually are crossed at discretion and drivers tend to watch out for swerving cyclists, if only for liability reasons.

This legislation is rather complicated but was imported here by the kind gentlemen who came to visit us in May 1940 and brought most of their traffic rules with them - which apparently suited us well.

In many of David's videos you see him taking the cycle path, even though he often has the choice of riding the main road instead (max 70 km/h or less). His Mango is a three-wheeler, probably just under 75 cm wide, which means that he is exempt from the obligation to use narrow cycle paths. But he would be a fool to take to the main road when the cycle paths available on his commute are wide enough to allow him to maintain speed.

Freedom is good for cyclists, but also for cars and their drivers. And if separation means that there are less accidents, so be it.

As for the road re-structure, there is an example right here where I live (in Assen, like David). I live on a main road which needs to restructured and widened. It has two car lanes, two cycle lanes/paths (partly lanes, partly separate paths), and two sidewalks. Two rows of trees separate the sidewalks from the other traffic. A bus lane will be added plus a central separation between the car lanes, and one of the cycle paths will be two-way (the other one remains as it is). The trees will be moved to fit the new profile. Fortunately there is room to make these alterations, but two buildings project too far into the road - a bank and an office building. The solution is very Dutch: the bank will lose its "Schalterhalle" as people do their bank business over the internet anyway, the office building will give up an entire wing. And a petrol station along the road will be removed altogether as nobody needs a petrol station in the middle of town.

Slow Factory said...

Frits B: How do trees get "moved"? Anyway, thanks for all the details. It is clear that we have different opinions on the necessity of cars.

Frits B said...

@GIF: "Moved" may not be the obvious word for uprooting the present trees and planting new ones a few meters away. But as the old one are somewhat rickety and the new ones will be healthy young saplings, I think the move will be for the better.

As for the necessity of cars, they do have their practical sides. And as it looks like they will persist until there is no more oil left to burn, we might just try to make the best of our cohabitation. I gave up driving 30 years ago, don't even have a licence anymore, and have never missed it. But then I had the opportunity to choose the place where I live quite close to where I had to earn my money.

Anonymous said...

Moving trees is often quite literal though, they changed the road profile of the area access road here from cycle paths on both sides to a single wider path on one side and they brought in a huge tree scoop to move the trees.

The scoop grapples the tree, then pushes down spades hydraulically to move the tree, with soil and all, works quite well for such short distance moves.