Traffic Signal Pie: Partially Protected Turns

There was a photo of a drawing doing the rounds earlier this week on Twitter showing a crossroads in Cambridge where there was a debate about how to fit cycling into a situation where the side roads weren't wide enough for cycle tracks.

The official designs for the junction showed a cycle track with parallel crossings (where people walk and cycle next to each other, but in separate space) over the side roads. All well and good, but with people cycling to the right of people walking (in the direction of travel), left turns for cyclists would be banned which is pretty useless if you need to turn left.

The photograph above shows what I mean - you can see the parallel crossing over the side road, but you can just see the banned left turn sign at the bottom of the traffic signals which are for cycle traffic (full-sized and low level). In this case, left turning cycle traffic actually turns off before the crossing and pops out in the service road in the distance and so is a kind of 'free' left turn. A closer view of the signals is below.

So, how could we provide for left turns in this sort of situation but where there is no opportunity to turn people in advance and where there's no space to turn people behind the pedestrian crossing points because the side road isn't wide enough for cycle tracks?

I don't think that doing nothing is the answer because apart from being a cop-out, people moving under their own power are just going to turn left anyway; not because they are reckless criminals, it's because it's an unreasonable requirement. If we are providing for cycling, then we need to do it properly by anticipating the likely (and desired) behaviour and accommodating it.

We could use toucan crossings such as the ones outside Hatfield Station in Hertfordshire because essentially, there is nothing technically illegal about turning left out of a toucan crossing; but, we are sticking cycling into walking space which is a fudge.

I think the answer might be what I will dub the "partially-protected left turn" (because I don't know if it has an official name); the protection being a combination of physical measures and clever user of traffic signals. Compare this idea with a fully protected left turn where a cycle track would continue into the side road to permit fully protected left turns.

The photograph above shows a Dutch junction which I've mirrored to show an approximation of what a UK protected left turn would be - mainly because I don't think I have a decent UK example in my photo library! The green arrow simply shows that the left turn can be made without worrying about general traffic - it's also called a free left turn (from the UK point of view).

So, here is my partially-protected left turn (above). The main road runs north-south in this example, with the side roads forming the crossroads east-west. As on the photograph above, we've one-way cycle tracks which cross the side roads on parallel crossings (which is key), but the difference from the parallel crossing photograph above is left turns for cycle traffic are catered for. In fact, right turns are also catered for because with this particular layout, the parallel crossings running together allow it; not forgetting that pedestrian crossings are also running at the same time on the north and south main road arms. Perhaps with the right turns, some further guidance markings might assist.

In truth, there is still some fudge about this design because the idea is that cyclists turning left or right from the main road cycle tracks (the north and south arms) will have to stop at a second stop line in the side roads immediately before the pedestrian crossing, depicted by the red circle in the image below;

This stop line would be accompanied by a low level cycle signal (LLCS) to be in the eye line of a turning cyclist. Once the parallel crossings and crossings of the main road end (all together, remember), the the LLCS on both side roads will go green and the turns into the side roads are complete.

In fact, we can be even more clever. If the LLCS from the partially-protected turns run before the side road traffic is released (using what is called 'early start') then the turns are completed way before other traffic gets into the side roads from behind them. Early start would also allow people cycling out of the side roads to turn left from the advanced stop lines (ASLs).

The other important feature is that in order to allow cyclists to turn right from the side roads, we really need to be releasing the side roads on their own stages otherwise cyclists potentially have to wait for a gap in oncoming traffic. It isn't vital we do this from a traffic signalling point of view, or even from a 'conventional' safety point of view, but it is a lower collision risk and would feel so much safer. Ideally, the side roads will be coming from low-traffic areas and so the green time needn't be too long which would allow them to run separately.

The sequence above shows how some of the signal stages would work. I've not bothered with the north-south general traffic stage; so starting with the parallel crossings/ main road crossings, you can see left and right turning cycle traffic stopping in the side roads at the stop line. Right turning cyclists would give way to oncoming cyclists as they would if this was general traffic running together.

Next, the green comes in for the road to the east. This allows the cyclists waiting in the entrance to the road on the east to proceed west and complete their turn from the previous stage. Cyclists coming from the eastern arm can turn left onto the cycle track to head south or to turn right to the far side to go on the other cycle track to head north.

There is also a green on the LLCS for the entrance to the side road on the east because there may be cyclists from the previous stage who are completing their turn - I've highlighted this in yellow (but it's a green signal). The last stage is just a reversal with the western arm getting a green, but we don't have the LLCS for previously turned cyclists as we did in the previous stage because there wouldn't be anyone left waiting (yes, it's complicated).

If the two side roads ran together, then the movements could be as I described, but the LLCS on both side roads would come on together, definitely with an early start to allow cyclists who turned on the previous stage to complete their movement. This would then essentially run as a 3-stage junction. With the stop lines in the side roads, there is a risk that some cyclists won't stop (don't write in) - we are dealing with people after all. However, is this any riskier than the toucan fudge or people ignoring a banned left turn on a parallel crossing?

I appreciate that this is fiendishly complicated to understand - it was to write down and I'm hoping I've got it right myself. I'm sure someone is going to spot a flaw which means I end up either rewriting or trashing this post (joke - hopefully).

Now, this isn't pie in the sky. There is a similar layout to this on the CS2 cycleway in East London. The junction of the Whitechapel High Street/ Commerical Road/ Leman Street has this layout. Well sort of, because Leman Street is one way going away from the junction. I think there are other similar junctions on CS2, but this one does show LLCS in the way I have described them, but to be fair, I should probably go for a better look and make a video!

The junction is much larger than in my example because left turning drivers are held when cycle traffic and ahead moving general traffic runs together (called 'hold the left turn'). Where cyclists get a green, they are permitted to turn left or move ahead; the right turns for cycle traffic is banned because because of the way the staging is set up (I'll come back to that in a minute).

The top image (above) shows the eastbound cycle track on Whitechapel High Street. As luck would have it, cycle traffic has a green on both full-sized signals and LLCS (circled green). In the side road (Commercial Street) you can just see a stop line and a red LLCS. Left turning cyclists would have to stop at the red as with my drawing above and the bottom panel shows a view once one is around the corner.

Right turning cyclists here would actually move ahead to the far side of the junction and then wait on the nearside to complete the right turn as a two-stage right turn - I'm not suggesting this at all in my layout, although you could do it - although the first time I used one in the UK, it went a bit wrong! The LLCS green to clear previously turned cyclists is still a two stage right turn, but you turn right before the second stage without having traffic thundering past you. In my view, this is far superior because you will be looking at a LLCS right next to you rather than a signal on the other side of the junction.

As well as the concept from a signalling point of view, there are some other features which are kind of nice to have and they can be applied in other places. First, we have the flare;

The cycle tracks approaching the junction are under some space constraints - whether this is actual width, a desire for a buffer (with planting) or some other issue. However, because of the more open geometry of the junction compared to the roads before the junction, we actually have a little more space to play with. The flare allows people to queue side by side at the stop line which increases throughput of cycle traffic. People turning right can also move to the right in readiness.

The left turn also flares a bit to help slower left turners to get out of the ahead/ right cycle traffic and the radius of the kerb here is designed to be comfortable for people using non-standard and adapted cycles. The other features in the close-up image above is the point that the cycle track surface ties into the carriageway surface - no kerbs to fail and throw the rider. The dotted white line is simply there to help show drivers the edge of carriageway. There is also the 'mini-zebra' which keeps priority with pedestrians who cross to the 'floating' crossing point for the signalised crossing of the carriageway.

The image above shows a set of the "elephants" feet which are designed to help show the route through a junction for people cycling. The markings are not parallel, which is a reflection of people being able to occupy the wider area before the stop line and then fall in tighter as they move through the junction and back onto the cycle track. As people move off at different speeds, they will naturally filter in.

The image also shows one of the disadvantages in that the pedestrian crossings are off the main walking desire line and the arrangement also means there couldn't be diagonal pedestrian crossings if walking and cycling run together in the same traffic stage.

Now, the layout can be tightened up for motor traffic because we don't design for the largest lorry which theoretically might turn into a side street and so with a bit more development, the pedestrian crossing points could be pushed closer to the walking desire line.

I think the problem we have in the UK is the lack of national design guidance. I know there is some good local guidance out there, but even there, we don't get enough example layouts on how we can deal with common problems. In fact, the reason this layout has made it onto my blog has come from a request for help from a campaigner on how the left turns at parallel crossings could be dealt with and the layout developed from a bit of back and forth discussion. I also tweeted out a request for an example of the arrangement at Whitechapel High Street because I was convinced I had seen one.

Obviously, I am interested in this stuff and so we now have a fairly accurate schematic and thought process (I hope) which solves the problem, but you are not going to find this in any design manual in this country. This is a problem because where we have designers working on projects who may not have the interest or knowledge in what is still a pioneering design subject in the UK (in a modern sense), we will struggle to think beyond what we are used to.

Engineers are very good at copying (what works or doesn't work to be honest) and so that is why we still see some really poor designs because they are applying layouts in old (and often discredited) national guidance. The government always tells us that local authorities are best placed to develop their own guidance, but that is nonsense because most don't have the resources or knowledge. We still need a big book of layouts and case studies. Anyway, have fun thinking about this one and feel free to criticise, because the layout is still on paper and it's easier to change than the kerbs!

On The Cheap

Building cycling infrastructure is often touted as an expensive thing to do, but this would mainly apply to street retrofits where a significant part of the costs are sunk into undoing layouts which were designed for motor traffic.

This does apply to walking infrastructure, although probably not to the same degree because most people walk and don't give footways and crossings a second thought.

In very rough terms, on a greenfield site where the land is already owned and there isn't much additional work to be done, the broad, all-in costs for building new is as follows;

• Footways - £100 to 150 per square metre
• Cycle tracks - £150 to £200 per square metre
• Carriageways - £200 to £250 per square metre

This is very broad, is subject to wide variation and depends on all three being built together. If a footway and a cycle track is built next to a carriageway, then the assumption will be that the costs are lower because they will benefit from carriageway drainage and lighting, whereas with a standalone footway and/ or cycle track, some of the costs go up a bit because of drainage and lighting. 

It's also down to the structural design. Carriageways need to carry motor traffic whereas cycle tracks and footways don't, although in the Design Manual for Roads & Bridges, there is recognition that they sometimes take motor traffic (usually over-run). In other words, a cycle track which can in theory be mounted by a delivery truck, needs to take this into account (whether or not people should be parking a truck on it). 

If the design doesn't appropriately take this into account, even the over-run by a single HGV could actually cause damage. The photograph shows a cycle track which cunningly doubles up as a fire path. If it wasn't built to take occasional over-run by a fire engine (which could be full of water and so very heavy), then the first time it is used, damage could result. The photo below shows the same scheme being built, demonstrating some serious thicknesses of material going in.

The other issues is that of frost. In the UK, we often have soft soils such as silts where the first half a metre could be damaged by freezing temperatures. We also have clays which can shrink and expand in dry and wet weather. 

We will always be building on the underlying soils and while the structural requirement for a footway or cycle track won't normally need to be too thick, the condition of the underlying soil might require us to dig more out from below. The point I am making is that building footways and cycle tracks require the same level of technical know how and site information as we would with a carriageway.

Where we are not building across a field, then we are definitely into reworking what we have. One useful principle is that it is easier to convert a carriageway into a cycle track or footway than the other way round. The main reason for this is utilities. 

The table above is from some industry guidance on the depths utilities should be laid - the full document can be downloaded here. Have a look at the last two columns which give depths within footways and carriageways. A cycle track would generally be treated as a footway in this case. Some utilities are the same depth, such as water which has to be out of the way of frost. Others have a different depth on the basis that those under a carriageway need protecting from traffic loading.

Telecommunications kit is one which can be a massive pain to deal with because there is so much of it and we have the differential depth. I'm often dealing with developers who need to build a new junction or access across a footway where a carriageway construction is needed and all of the utilities are too shallow. If there is optical fibre, then best you have plenty of space for zeros in the amount box of your cheque book.

The other thing to consider with a scheme is that getting utilities diverted away from your scheme or lowered within your scheme is that it can take ages. On an important road or street, any work that will take over ten working days will generally need a 3 month notification period before it can start. This is to give proper notice to the utilities or others who might have works which need coordination (and could be a blog post in its own right). 

Street lighting is a classic item that takes ages to shift because of the need to disconnect, shift and reconnect. Traditionally the power works are undertaken by the power company and the column works by the local authority, but thankfully we now have contractors who are permitted to do both which makes coordination a bit easier.

The photograph above shows a trench being cut into the carriageway for a new kerb line for the interface between a cycle track and narrowed carriageway. This generally doesn't affect utilities (unless a chamber is in the way) and as the levels are being raised, we won't have to think about lowering utilities. In the situation where we have a very wide footway and we are taking some of it for a cycle track, even building a stepped cycle track (the cycle track lower than the footway, but higher than the carriageway) may reduce the cover to utilities enough for them needing lowering.

The photo above shows the southern side of Vauxhall Bridge in London. The footway and cycle track are at the same level (with a tactile kerb between) because the cycle track has half come from the carriageway and half from a wide footway. Dropping the footway half to carriageway level here would probably have had an impact on utilities. Further north, the cycle track as at carriageway level because the cycle track fully takes space from the carriageway.

The pair of photographs above are on the western side of Westminster Bridge in London. They are actually showing two different bus stops (one after another), but again, the cycling space and floating bus stop comes from the carriageway and so the cycle track can be kept at carriageway level. 

Part of the work included resurfacing the main carriageway after the works and a rebuild of the central reservation, neither of which were required for the cycle track. In the event of someone looking at the project costs from a criticism of investing in cycling point of view, all of these costs will be lumped in and so this really can skew how people see the costs of just providing for cycling. 

In my view, renewing the whole street is often a good idea from a maintenance point of view, but we need to be honest about where the costs are apportioned. 

The other piece of street kit to talk about is that of traffic signals. Costs vary with site of course, but replacing the equipment on a simple 4-way signalised junction might cost around £60k. Installing the same kit first time at the same junction might cost over double that because of the ducting, installing tactile paving and possibly moving lighting or other utilities round.

What we should remember with traffic signals is that the are there in order to manage the dominant flow of motor traffic. They give drivers coming from side roads a chance to get on the main road, or they protect people walking and cycling.

The photo above shows the junction of Ruckholt Road and Orient Way in Waltham Forest. Without the traffic signals, nobody cycling is going to have a fun time trying to get across the main road. The signals are not there because a cycling scheme was built per se, they are there because of the motor traffic and so the investment in the scheme is skewed towards dealing with motor traffic.

Finally, lets compare the costs of a zebra crossing over a carriageway and a cycle track. The first is perhaps £15k+ and the second is perhaps £5k. They will both have tactile paving and dropped kerbs in common, but the cycle track doesn't have to have Belisha beacons and it doesn't need high grip road surfacing to help speeding drivers slow down. Again, the extra cost is because of the motorised traffic and skews the budgets accordingly.

So, despite what some people might have you believe, building for walking and cycling is cheap, it's just that at the same time we have to blow a load of cash to accommodate motor traffic in doing so.

Crash Friendly

We are concerned that nobody is hurt on our streets and roads as a result of a mistake by them or others and this underpins a better approach to highway design. Of course, when someone gets it wrong, they shouldn't pay for it with their life.

Despite the promises of self-driving vehicles always being ten years away, we have to live in the now and that means we have to install 'stuff' into the highway to warn/ inform/ regulate people and that means traffic signs, bollards and the other kit used to manage the flow of people. The trouble is, people do make mistakes and the kit tends to be crashed into. This could be anything from the almost comedic person walking into a lighting column because they were distracted, to a driver on a high speed road having a blow-out and crashing into a bridge pier.

Perhaps we can think about energy. On the one hand, walking into a lighting column can hurt, but hopefully, you wouldn't be too damaged; although some people are less able to recover. On the other hand, crashing a car at 70mph into a bridge pier is almost certainly going to kill everyone in the vehicle. In other words, the greater the energy involved, the greater the risk.

For the lighting column example or indeed any piece of street furniture, we should take care with placement so that we don't leave it in a desire where someone might fall over it or walk into it; visually impaired people are going to be disproportionately affected. For the bridge pier example, we are going to have to put in some sort of protection which will absorb the energy of the vehicle and either stop it hitting the pier or divert it away.

One feature of our roads (motorways and trunk roads in this case) are traffic signs. Because of the speed of traffic, the text on the signs is large and so the signs are also large. Large signs present themselves to the wind and so they end up being attached to chunky posts with substantial foundations. You really don't want to crash into one of these large posts as the car tends to wrap itself around the post.

In some situations, safety fence (popularly known as crash barrier) is rolled out. You've all seen the type of thing, here's a video of it being installed;

The system uses posts installed into the ground with the barrier attached. In the event of a crash, the barrier is designed to catch the vehicle and absorb the energy. Some barrier systems are tensioned to further help with absorption. People often talk about wanting to see safety fencing along footways and cycle tracks next to main roads, but this doesn't mean that they'll be kept safe as this crash test shows;

The issue is that in order to absorb energy, the posts are going to snap and the barrier will buckle and stretch - you don't want to be standing behind it. This is also the reason that if you break down on a motorway, don't sit on the barrier waiting for help - try and get away from the road as far as you can safely go. Safety fence design considers this movement called the 'working width' and so any paths behind need to be set back beyond this working width. Of course, getting people away from the very big roads would be a better answer, if not always practical.

The problem is, we cannot put safety fence everywhere. It's costly, it requires maintenance and replacement and in some places, it's just not practical because of space or other issues. We can take a risk approach and so where we have bridges and gantry signs, safety fence is useful. For large signs, we have other options - passive safety posts.

The photo above shows a very large sign next the A13 in London where the speed limit is posted at 50mph. A sign of this size needs some chunky posts to hold it up, but the available space means it's quite close to the road and in the event someone loses control, they could well crash into the sign and its posts. The designer has chosen to use 'Lattix' posts which are strong in use, but in the event of an impact, they crumple;

We also have other things to contend with such as lighting columns. On trunk roads and motorways, they tend to be quite large and again, it's not practical to have safety fence in front of all of them. A risk based approach can be taken and as with sign posts, there are energy absorbing lighting columns which fold and buckle as they absorb the energy;

Motorways and perhaps to a lesser extent, trunk roads, are very consistent in layout and opposing traffic flows are physically separated. This hasn't happened by coincidence, it is the result of work done to develop standards for high speed situations which are mandatory and require sign-off processes to depart from those standards. Of course, once you throw people into the mix, things can be a little unpredictable and very dangerous for the people working there;

Cycle Tracks Should Be Laid In Red Asphalt

Ever since my first visit to the Netherlands in the summer of 2015, I have been obsessed with the use of red asphalt for cycle tracks (OK, the Dutch do design quite well generally).

The main reasons are that using a coloured surface helps to provide visual priority in situations such as cycle tracks crossing side roads and legibility in terms of people cycling being able to see how their route is laid out ahead. 

Here's the first cycle track I closely inspected in Deventer in 2015. It is of course red and in fact, each area of each mode of transport has a different surface which helps with legibility. The grey slabs provides the footway, the brown blocks is a hard strip which provides protection from traffic in the photo and behind me, it becomes parking bays and access points to parking areas and houses.

Here's a view a but further along - as turns out, the brick paving in the carriageway has since been changed to asphalt and some of the road layout has changed since I was there; such is the Dutch approach of constantly renewing and tweaking their streets.

It doesn't really matter where you go, there is a pretty good consistency with newer layouts tending to go for the black asphalt for roads, grey for footways and red for cycle tracks (give or take local approaches). Above is from Amsterdam and below is from Maastricht.

It's a pretty simple palette to be sure, but it is consistent and in my view gives a little more interest than surfacing everything in black asphalt which is so beloved of UK engineers. Below is a photo from Utrecht and below that a photo from Harderwijk.

Of course, you can use any colour you like and a greater part of the wonder of cycling in the Netherlands is the design which goes into the layouts, but I think they would be diminished without the red cycle tracks. 

You'll also notice that the Amsterdam and Maastricht examples are very bright, but that is because they are new. After some time, the colour fades a little and in my view, you are left with a warm-looking finish - there's just something about it I like.

Just as important as the colour is the fact that the surfaces are laid by machine so that they are smooth. In some part of the the country where the infrastructure is a little older, you will find small red concrete slabs on the cycle tracks, but these tend to get changed during street renewals such as the example below from suburban Amsterdam.

You can of course use any colour you like. In some parts of the UK, green has been all the rage such as here in the City of Manchester;

London has seen extensive use of blue (although it seems less in favour now);

In the UK, there is no rules about what colour you can use, it is up to each highway authority. In fact, you don't have to use coloured asphalt. It does bring me onto the costs involved. 

Black is often the default because it's the basic colour of basic asphalt and so it's the cheapest. In the world of austerity, there is a good argument for spending the money on getting a safe layout and the colour doesn't really matter, but I would argue that a coloured surface does play a part in it as I suggested that the start.

Red is the most common colour available after black and many asphalt plants keep a bin of red AC6 on the go because it does get used around the place where people though red footways were a little posher for their conservation area and other places where people wished to keep up appearances. Of course, where green is use a lot, then asphalt plants might keep that as a standard, although it's less common in my experience. Blue and other exotic colours are far more specialist and proprietary; frankly, I'd steer clear of them because they are so more expensive.

Maintenance types would have us use black asphalt everywhere because it is cheap and I get that. There are worries about having to repair coloured surfacing because small loads for patching work becomes expensive. But actually, it doesn't matter because cycle traffic doesn't destroy the surface and if small repairs are needed, then patch them in black. Eventually, if you end up with lots of patches after many years, you'll be resurfacing anyway.

So let's spend a little on a nice surface. Let's roll out the red carpet for cycling; and as well as the care and attention we have given to the space, let's help reinforce cycle traffic as a separate mode in its own right and as Catriona says, red surfacing is awesome!

Salford. Mmmm red.