Why do riders crash in corners?

Why do riders crash in corners? After all, had they not successfully negotiated a lot of other corners before they reached the one that got them? Was there something about certain corners that made them more likely to be accident sites? Why did these riders select a particular corner entry speed that proved to be so spectacularly incorrect? What is the process that we use for judging the severity of a corner and selecting a suitable entry speed? Do we all use the same method, or are there a number of ways in which we can analyse a corner before we reach it?

Some years ago a team of scientists wanted to answer these questions, so they wired up some drivers with a device that overlays the point of a drivers gaze onto a video, so that they could see where they were looking as they entered and negotiated a corner. Their findings were quite interesting as it seemed that their initial gaze was directed towards something called the Tangent Point (TP) which is the point of the inner lane boundary bearing the most curvature on the retina. Looking at a 2D picture of a road, we see that this point of maximum curvature forms a hook in the kerb.

Tangent point

The scientists found that most of their test drivers naturally fixated on the TP as they negotiated a bend and in fact about 75% of their time was spent looking at this point rather than any other point on the road. They also found that drivers on their way into a corner glanced at the TP about two to three seconds before turning the wheel and then reacquired it at around half a second before making a steering input. The degree of hook at the tangent point will give a fair estimation of the severity of the upcoming bend and the rate at which it is coming towards a driver will give a good indication of by how much they might need to reduce their speed.

It seems then that the degree of curvature at the Tangent Point is the crucial piece of information we all use to set ourselves up for a corner. If this is the case, then if we get the corner wrong, we must have misinterpreted the curvature at the Tangent Point. This then is where the left hand bend finally begins to offer up its secrets.

Most of us think of a corner as being a constant curve linking two straights going in different directions. For most of the time, we would be right in this assumption, but not always. We now need to study a bit of history and strangely not of road transport, but of the railways. When speeds on the railways were very slow, straights were indeed connected together by constant curvature bends as this was by far the easiest way of planning the layout of the tracks. It soon became apparent that as speeds increased, the passengers were getting thrown to one side of the carriage as the train went directly from the straight onto the curve. This sudden onset of ‘G’ in the corner was considered to be a bad thing for the passengers, so the railway companies started to look for a way round the problem. They soon came up with the idea of the spiral transition, or clothoid easement curve. The easement curve as its name suggests, eases the carriage from going straight to going round the ultimate curve. By using this track layout, the railway companies discovered that the amount of ‘G’ could be increased gradually, thus allowing the passengers to feel its gentle onset and brace against it so no more spilt tea and passengers heaping up on top of each other in bends, a perfect result!

All would have been well had the spiral transition or easement curve stayed where it was on the railways, but somehow it escaped. Those readers who have done an advanced riding course will be well aware of the quest to get riders to turn the bike quickly. Turning quickly is a very good thing to do, but no matter how quickly you turn a bike or a car, the path you take from going straight ahead to turning will ALWAYS be a spiral. If you can imagine turning slowly, then at the beginning of the turn the bike will not be banked over very far so it describes quite a wide arc. As the angle of bank increases, the radius of the arc gets smaller and smaller until it matches the arc we need to negotiate the corner, this then is a classic spiral. We can’t help but do this as it’s the laws of physics working away unnoticed and unheeded.

Back in the mists of time a chap called Henry Criswell who was the county surveyor of Devon, decided that in order to promote passenger comfort; roads should be designed with easement or transition spirals installed at the beginning of bends. His book “Highway Spirals, Banking and Vertical Curves” Became the seminal work for all highway design and construction engineers and is still in use today. What he singularly failed to realise at the time was that the driver already had control over the rate of spiralling into a bend by dint of their ability to steer the vehicle, which is something that a railway train, constrained by the track, simply cannot do.

Thus the spiral easement became part of our modern roads landscape and everybody went on their way rejoicing in improved passenger comfort. Except that it didn’t improve passenger comfort as that was under the control of the driver not the road. What it did do was to fool riders and drivers into the possibility of making a very serious error. Let’s hark back to the beginning of this little piece when we discovered that the rate of curvature at the Tangent Point was critical in our understanding of the severity of an upcoming corner. The clothoid spiral easement throws an enormous spanner in the works by showing a radius of curvature that is far less severe than the radius of the actual curve!

spiral-circle corner

If we judge the corner by the curvature at the Tangent Point and the curvature is not that of the actual corner, we are going to get it wrong. If we get a left hand bend wrong, then we will run wide into the oncoming traffic, which unless we are very lucky, will kill us stone dead.

Is this then the secret of the single vehicle accident on rural left hand bends? Many years ago, back in 1977, a young engineer called Doug Stewart who was working for Aberdeenshire Highways Department noticed a significant number of accidents were occurring on just a few bends in the County. Being of an enquiring mind, he wanted to know what the difference was between these bends and other equally severe bends that had noticeably fewer accidents. The bends in question, quite naturally featured our old friend the easement curve or spiral transition. As a Highways Engineer, he was lucky when one of the killer curves was scheduled for re-surfacing. He persuaded the Council to re-align the curve from a spiral to a circular curve so that they could see if the realignment made any difference to the accident statistics. As you might have guessed, the number of accidents fell away significantly after the change.

So now we have identified the problem what can we do about it?

With the best will in the world it would take forever to realign every spiral transition in the country and that is supposing that the powers that be admit that they are a significant a problem in the first place. Once again it is down to us riders to work out a strategy for handling transitional spirals so that we can identify and negotiate them without running wide. First thing I did when I learnt about these horrid things was to go out and see if I could find any in the local area. That proved to be an extremely difficult task as the easement curves are very well disguised, which I suppose is the reason we have the problem in the first place! I could only identify a transitional spiral for certain after I had gone round it and not before. The key identifier of a transitional spiral is the fact that you have to make two steering inputs in order to negotiate it. The first carries you round the spiral and the second to carry you round the actual corner. This, sadly, is thruppeny bit-ing, yet unless we do it, we cannot safely negotiate a bend with a spiral transition, this is a terrible conundrum, but it is the lesser of two evils, either steer once and run wide or steer twice and make it round.

As we can’t tell by looking whether or not a corner has a spiral transition, we must be prepared to make two steering inputs at any corner, especially those we are unfamiliar with. Being prepared in this way means that we don’t have to work out a strategy on the fly, we can simply go to a pre-planned response once we discover the true nature of the bend. This does mean that we have to keep a fair bit of bank in the bank, but all riders should have no problem with doing that.



15 thoughts on “Why do riders crash in corners?

  1. I’m afraid I persist in believing in the ‘Limit Point’ or the vanishing point… I find that while this is moving towards me, I’m not going to make my attack on the corner. Of course, I may have technically ‘turned in to a bend’ if it spends a long time gently tightening, because you have to turn if you are to stay on your own side of the road, but I haven’t <>… only as the corner starts to open out, as the limit point starts to recede, am I moved to commit slightly harder, and only when it shoots off spectacularly into the distance am I likely to properly commit to the corner. I’m interested in the research that shows that most drivers don’t do this – I’m not sure I did it back when I was purely a car driver, it’s a learned behaviour, but as a strategy for reading {flat} corners and responding to their variations in profile, it works reliably enough for me I think…


    • There is a great deal of research that has been carried out since Land & Lee’s original work in the early 90’s on steering strategies (just Google “tangent point land lee” and you will find a great many papers on the subject. This 2010 paper from Kandil, Rotter & Lappe represents the current state of understanding about the steering process. http://www.journalofvision.org/content/10/4/24 and covers the differences in strategies between ‘closed’ and ‘open’ corners.


  2. I must be strange, because I don’t spend much time looking at the ‘TP’ marked on your picture, except to check for the state of the road surface. Once I’ve clocked that, I stare hard at the vanishing point instead, which is further down the road where the surface disappears out of sight. If the vanishing point isn’t moving away I know that the bend is tightening up so I lose speed. The other advantage of staring at the vanishing point is that it tells you if something big is coming the other way! I had no idea that most people look at the TP, no wonder things come as a surprise.


  3. I wonder if Henry Criswell should be added to the Hall of Fame of worst inventors in history! Along with Thomas Midgley Jr who invented Leaded Petrol and CFCs:
    “Midgley had more impact on the atmosphere than any other single organism in Earth’s history.”


    • The spiral seems to be something unique to the UK, New Zealand and the Parkway’s of the USA. It would be interesting to find out about corners on the Continent as that might provide us with some ammunition.


      • I do not agree with the article, austrias roads are built with combination of transisting spirals (Clothoides, Cornu curve) and circular centerparts of the corners.

        A car draws a clothoide more or less when turning the wheel with constant angular speed.

        The earlyer you start to move the steering wheel, the smoother the lateral acceleration increasing linear, not exponential.

        Problem is not the lateral acceleration that will bring you from static friction to dynamic friction, it is lurch.

        Lot of car drivers start steering too late, an then they get victims of centrifugal force at a point, when lurching into dynamic friction.

        On spiral corners, you will be able to keep your car on the same side position, without any side movement within your lane.
        Interesting for ne are two facts:

        1) There seem to be differences in the dynamic when cornering with a motorbike.

        2) There might be other problems in the left corners, caused by our patterns of eye movement, that are different in viewing from left to the richt / Right to the left. Maybe this is the effect, when some are fixing the tangential point in left bends, just an idea.

        Circular bends maybe better for Motocyclistst, with quite other degrees of freedom in possitioning their vehicle.


      • The spiral easement certainly ‘works’ for cars Fritz, but only if they stay below the design speed. Above that and it begins to feel very unpleasant which is why they are favoured for the design of Parkway roads in the USA. The way a motorcycle is steered makes the spiral easement far more of a problem for bikes as everything starts to go wrong as soon as the design speed is approached let alone exceeded. They require two steering actions on a bike whereas they only require one in a car and we believe it is that which makes them a problem.

        Much more research is required on this particular problem, but sadly the authorities refuse to recognise it as a problem as they think it’s always down to the rider going too fast! Circular bends are much easier for a rider because they do not disguise their true nature which means they do not force riders into making a speed selection error. If it is a speed selection error that the authorities are concerned about then it’s surprising they don’t care to find out why such errors are made in the first place.


  4. Sometimes that gradually decreasing radius is just the shape of the terrain Mother Nature gave the roadway engineers to work in.

    In the fairly large, mostly rural county I live in, two curves have taken two lives each over the past 10 years. That’s out of many hundreds of curves over hundreds of miles of roads and an average of 5 motorcyclist deaths per year in all riding environments.

    Both curves have that shape, but it’s merely an accident of the surroundings. Viewing a motorcycle crash map built by a university traffic research group, the two stand out among many nearby crash-free bends.


    • The trouble is Dan that we have far too little understanding of the prediction failures that lead up to accidents on corners. Even though they account for around half of the fatalities in the UK there has been not one single study or any research done on the actual cornering process. It doesn’t seem to matther whether or not a bend naturally tightens or features a spiral easement it seems like riders are getting fooled by them. I attended a Government presentation recently that also featured a crash map that showed a significant cluster of accidents on just one bend. My colleague commented that he knew the corner very well as it was noted for the way in which it tightened without any prior indication. The answer seems to be that if a spiralling or tightening bend is made easier to identify then fewer people will crash at them.


  5. Duncan, the design speed should not be that problem, a self explaining road design wiil give you scope enough. spiral or circular, over design speed wil make you cutting the bend, within the lane, a biker has some extra space for a larger radius, but then beeing fixed on this “fast line”.

    Think we have to make firdt a statement wich “line” is accordin to rules of the road: “a comfort line (time and space to adjust) ” , “the ideal line- smooth and active with some reserve” or a “hard competition line – at the ege of physics, without any extra escaping features.”.


    • The best ‘escaping feature’ is always going to be the ability to tighten an established line (make another steering input). If a rider takes a hard racing line then they sacrifice any ability to ride the corner any tighter than they originally planned to ride it. In every corner on the roads at least, a rider should have an ability to tighten their pre-selected line should anything happen that would require it. The idea is to ‘keep some bank in the bank’ so that the line can be tightened if neccessary. This also requires significant amount of practice in tightening lines mid-corner otherwise the bike may be capable of turning tighter, but the rider will not be.


  6. Pingback: Impress yer mates (if you have any) #1: Clothoid Curves | pothole surfer

Something to say? Leave a Reply.

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s