Science of cycling still largely mysterious. No one knows why bikes stay upright

Enkidu

10 mW
Joined
Jan 4, 2022
Messages
28
Location
Southeast Michigan, USA
I found these very interesting.

Science of cycling still largely mysterious; CBC News "Their basic mechanics are understood, but there are many questions about the physics of bikes. "It's as simple as riding a bicycle" is a common expression. But the science of staying upright on two wheels is anything but simple — and we know surprisingly little about the intricacies of how cycling actually works."
https://www.cbc.ca/news/science/science-of-cycling-still-mysterious-1.3699012

The bicycle problem that nearly broke mathematics; https://www.nature.com/articles/535338a

How Do Bikes Stay Up? https://www.youtube.com/watch?v=oZAc5t2lkvo

We still don’t really know how bicycles work. Forget mysterious dark matter and the inexplicable accelerating expansion of the universe; the bicycle represents a far more embarrassing hole in the accomplishments of physics.
https://www.newstatesman.com/culture/sport/2013/08/we-still-don%E2%80%99t-really-know-how-bicycles-work

Your bike’s secret to staying upright is actually a mystery. Research project suggests it’s A Lot More Complicated than thought. https://www.bikeradar.com/features/your-bikes-secret-to-staying-upright-is-actually-a-mystery/
 
I've always wondered if the solution to this mystery might also somehow reconcile the discrepancies between Newtonian and Quantum physics. At first glance, the problem appears to be described by Newtonian physics, except no one has to date been able to come up with an explanation that is congruent with such, so maybe there are non-Newtonian components to a model that would explain how a bicycle stays upright?
 
Lol, no. Quantum physics has nothing to do with it, and it is easy to make a robot that rides a bike - just steer into undesired lean and 'preplean' (countersteer) to initiate a turn.

*Self-stability* is a bit more complex, but effects pretty subtle and bike lacking it is still perfectly rideable.

The general question is 'what makes a bit feel *effortless*" to ride to a HUMAN, and which nuances of frame/steering geometry make for a best handling bike.

First, that's not a question with a single answer, because a bike that is perfect for one scenario will be poor for an other, and the most important - is has less to do with bike itself, but with bike-human INTERFACE, and the human in question usually has absolutely no idea what he is doing to actually ride a bike - riding a bike less of a skill and more of a *habit*. A lot of people are not simply ignorant of, but vehemently deny that they are doing any 'countersteering'!

So, suddenly we have not a problem of a bicycle with two wheels, frame and handlebars, but that bike AND human body on top that is astoundely complex arrangement of levers and actuators with tons of degress of freedom, and a black box control system that needs 'reverse engineering' to solve the 'interfacing' problem!
 
Once you can track stand indefinitely, the gyro stability myth carries little water.

I think it's the human driving the wheel contact patches to always stay under the dynamic center of gravity as the bicycle turns.
 
About 50 years ago they wrote stories that it was scientifically impossible for helicopter to fly ... fortunately the helicopter pilots were smart enough not read those stories.
 
BalorNG said:
Lol, no. Quantum physics has nothing to do with it, and it is easy to make a robot that rides a bike - just steer into undesired lean and 'preplean' (countersteer) to initiate a turn.

I knew that would garner laughs. The point being, to look for an explanation in unexpected/unlikely/impossible places since the conventional approach has yielded no explanation.

liveforphysics said:
I think it's the human driving the wheel contact patches to always stay under the dynamic center of gravity as the bicycle turns.

How would one set up an experiment to prove or disprove that?
 
I flew today, it was quite the ride :mrgreen:

999zip999 said:
Yeah you must question me about levitating but ask LFP about flying. And in some places it's legit.
 
The Toecutter said:
BalorNG said:
Lol, no. Quantum physics has nothing to do with it, and it is easy to make a robot that rides a bike - just steer into undesired lean and 'preplean' (countersteer) to initiate a turn.

I knew that would garner laughs. The point being, to look for an explanation in unexpected/unlikely/impossible places since the conventional approach has yielded no explanation.

liveforphysics said:
I think it's the human driving the wheel contact patches to always stay under the dynamic center of gravity as the bicycle turns.

How would one set up an experiment to prove or disprove that?

You mistake problem of riding a bike with a hard problem of counsciousness.

It's an easy mistake to make, because they are indeed closely connected - for reasons I've stated above (we have counscious *illusions* of knowledge that we *know* how to ride a bike, despite having zero, only a *habit* of riding it using brain subsystems that are isolated from counscious awareness). Yet, there is absolutely nothing mysterious why bikes stay *upright* with human input and even self-stable without it (that require a range of conditions though). It is putting it into a neat set of equations that are true for all conditions is that is hard.

However, 'hard problem of counsciousness' (how brain activity creates *feeling* of being you) is a different problem and we indeed have little idea how it is *actually* accomplished, and there are (plausible) theories that quantum mechanics have something to do with it... but again, overall idea that your brain creates it somehow is basically rock-solid science - neuroscience, it is details that are elusive because brain is literally the most complex object in the known universe, quantum mechanics or not. When it comes to 'quantum consciousnesss' come neuroscientists joke that entire logic of brinking former into it verges on arguments like: 'quantum theory is complex and mysterious, consciousness is complex and mysterious, therefore they must be related somehow!'.
 
So we can stabilise fighter jets, fly quadcopters up down left right... down a chimney and out the door, we can land rockets on the moon, stabilise orbits around our planet, slingshot probes off other planets, stabilise flows of subatomic particles close to the speed of light in 27km tunnels we've dug under mountains, steer drill strings 10km below the surface of the earth... But we don't know how bicycles stay upright?

I think a few uni profs and researchers need to put down their crack pipes and try a few lines of math. Or maybe I'm a world expert in bike mechanics and should get a Nobel prize.

https://m.youtube.com/watch?v=B6vr1x6KDaY
Here's something impressive in the world of control.
 
JackFlorey said:
mxlemming said:
But we don't know how bicycles stay upright?

Of course we do. The articles themselves say that we know. The thread title is just clickbait.

:lol: and we're meant to trust CBC and nature...
No my contention and dismay is that there are genuinely still people getting paid to research this at university posts. Click bait or not.
 
mxlemming said:
No my contention and dismay is that there are genuinely still people getting paid to research this at university posts.
I am all for researching bicycle dynamics. But that's because I think we should learn more about how they work and how to improve them - not because "no one knows how they stay upright."
 
Training an algorythm is a working knowledge, which is undiminished by having 0% understanding of how it works.

I've seen enough videos of bears and primates riding bicycles or motorcycles with effective control. I doubt the guy teaching them starts with equations about gyroscopes and center of gravity relative to tire contact patch points and steering vectors.

The guy programming the robot makes the best model he can, and then PID loops and high speed decisions based on tuning the rates and loops keeps it up.
 
What bit about this exactly do "we" not understand? It seems pretty obvious to me. When the bike is rolling, a deviation from upright means the weight of the bike causes the steering to twist which in turn creates a corrective centrifugal acceleration to stand the bike back up.

There's also effects from the gyroscopic wheel precession torques and the width of the tyres producing a torque to stand the bike back up but these are neither essential nor sufficient to fully stabilise the bike.

You can endlessly twiddle the exact geometry to get different gains on the lean angle vs restoring force produced, but there's a wide window that serves the purpose of keeping the bike upright.

At low speed, it requires active user input to balance, which is done intuitively by shifting the wheel position under the centre of mass by turning the front wheel.

What am I missing?
 
mxlemming said:
What bit about this exactly do "we" not understand? It seems pretty obvious to me. When the bike is rolling, a deviation from upright means the weight of the bike causes the steering to twist which in turn creates a corrective centrifugal acceleration to stand the bike back up.

There's also effects from the gyroscopic wheel precession torques and the width of the tyres producing a torque to stand the bike back up but these are neither essential nor sufficient to fully stabilise the bike.

You can endlessly twiddle the exact geometry to get different gains on the lean angle vs restoring force produced, but there's a wide window that serves the purpose of keeping the bike upright.

At low speed, it requires active user input to balance, which is done intuitively by shifting the wheel position under the centre of mass by turning the front wheel.

What am I missing?

There is also countersteering phenomena. And again, it is putting it into neat set of equations that is hard - pretty much like, say, turbulent aerodynamics. We have planes that fly nicely, but better understanding and better models might widen our 'design space' and *might* allow, say, better fuel economy - like 'shark skin texture' or something that emulate feathers... but doing it without true understanding how it works is just a cargo cult - *even* if it ultimately works.

When it comes to bicycles - same thing. Is there a model where I can, say, replace typical head tube steering with virtual pivot steering with very different (and non-linear!) relationships between steering angle, trail, return to center force and side force steering input, so I can have myself an fully faired HPV that is stable in windy (and turbulent windy!) conditions, that outputs a solution with maximized 'steering feel'?

Which steering interface (height, width, pivot angle of the bars) should I optimally choose? Which body position gives me best aerodynamics and best steering input in all conditions and allow me to ride at slow speed and in loose/rought conditions with impunity?

How to quantify 'steering feel', anyway? How much is enough to satisfy 'with impunity' condition?

Those are not abstract philosophical problems - those are engineering challenges I'm trying to solve right now, and if you have answers to that I've very much like to hear that.

Basically, you can only claim to understand a system when you can break it down to the most basic parts, assemble it a drastically different configuration and predict with certainty how it is going to behave - preferably in a somehow better way from original.
 
Anyway, as much as I understand this (I'm not a mathematician, unfortunately) - this is not unlike an n-body problem (where n is greater than 2), where once you add enough variables interacting in complex non-linear ways to the system, it stops having a 'solution' at all.
Still, when you formulate the question as 'why bikes stay upright without a rider' (self-stability) - yea, now we more or less have answers to this. But it has very little to do even with handsfree riding I suspect - for instance one of my recumbents has zero self-stability (zero trail, not weight stabilisation) and indeed does not allow me to ride handsfree - but otherwise the best handling recumbent I've ever ridden... maybe because it is not really a 'recumbent', having a very upright torso position and very wide bars that you can lean on - but underseat ones.
 
It also turns out that no one is really sure why planes stay in the air either.

https://www.scientificamerican.com/article/no-one-can-explain-why-planes-stay-in-the-air/#:~:text=The%20theory%20states%20that%20a,or%20flat%2C%20symmetrical%20or%20not.


BalorNG said:
mxlemming said:
What bit about this exactly do "we" not understand? It seems pretty obvious to me. When the bike is rolling, a deviation from upright means the weight of the bike causes the steering to twist which in turn creates a corrective centrifugal acceleration to stand the bike back up.

There's also effects from the gyroscopic wheel precession torques and the width of the tyres producing a torque to stand the bike back up but these are neither essential nor sufficient to fully stabilise the bike.

You can endlessly twiddle the exact geometry to get different gains on the lean angle vs restoring force produced, but there's a wide window that serves the purpose of keeping the bike upright.

At low speed, it requires active user input to balance, which is done intuitively by shifting the wheel position under the centre of mass by turning the front wheel.

What am I missing?

There is also countersteering phenomena. And again, it is putting it into neat set of equations that is hard - pretty much like, say, turbulent aerodynamics. We have planes that fly nicely, but better understanding and better models might widen our 'design space' and *might* allow, say, better fuel economy - like 'shark skin texture' or something that emulate feathers... but doing it without true understanding how it works is just a cargo cult - *even* if it ultimately works.

When it comes to bicycles - same thing. Is there a model where I can, say, replace typical head tube steering with virtual pivot steering with very different (and non-linear!) relationships between steering angle, trail, return to center force and side force steering input, so I can have myself an fully faired HPV that is stable in windy (and turbulent windy!) conditions, that outputs a solution with maximized 'steering feel'?

Which steering interface (height, width, pivot angle of the bars) should I optimally choose? Which body position gives me best aerodynamics and best steering input in all conditions and allow me to ride at slow speed and in loose/rought conditions with impunity?

How to quantify 'steering feel', anyway? How much is enough to satisfy 'with impunity' condition?

Those are not abstract philosophical problems - those are engineering challenges I'm trying to solve right now, and if you have answers to that I've very much like to hear that.

Basically, you can only claim to understand a system when you can break it down to the most basic parts, assemble it a drastically different configuration and predict with certainty how it is going to behave - preferably in a somehow better way from original.
 
Enkidu said:
It also turns out that no one is really sure why planes stay in the air either.

https://www.scientificamerican.com/article/no-one-can-explain-why-planes-stay-in-the-air/#:~:text=The%20theory%20states%20that%20a,or%20flat%2C%20symmetrical%20or%20not.

This article, while it does a good job of trying to push idea that aerodynamics is all about interacting FIELDS and makes no sense without temporal dimension (Von carman sheet is a great of example of this), does not mention thermal velocity of air molecules anywhere and this is key to true understanding of lift/drag and also why exactly 'speed of sound' is so damn important.
 
Back
Top