Nice piece on the difficultly of reprogramming your brain to steer in the opposite direction as normal

32 views
Skip to first unread message

Jason Moore

unread,
May 5, 2015, 5:30:01 PM5/5/15
to Single Track Vehicle Dynamics
This is from the youtube program "Smarter Every Day". The host spends a few months training himself to ride a bicycle that steers in the opposite direction as normal. After learning to ride the bike it is initially impossible for him to ride a normal bicycle anymore. His child learns to ride the bike very quickly (relatively speaking).

Enjoyable to watch if you've never seen or tried on of these bikes.

If riding this bike simply requires us to change the sign of our feeback gain(s), I wonder why that is so difficult to do?

kyle

unread,
May 6, 2015, 3:40:20 PM5/6/15
to st...@googlegroups.com
Hi Jason,

Fascinating video - interesting how much quicker his son was. I've
driven on the left hand side of the road for over 100,000 miles in
Australia, New Zealand and England. It is easy to switch back and
forth. I wonder if he will be able to ride both kinds of bike as
easily?

Chet


Jason Moore

unread,
May 6, 2015, 4:14:22 PM5/6/15
to Single Track Vehicle Dynamics
Andy Ruina came across this relevant paper:

http://www.jneurosci.org/content/14/5/3208.full.pdf

The authors found that if you change the dynamics of the system, in this case reaching with a manipulator, we will gradually update our internal "inverse dynamics" model such that we end up having the same trajectory as we did with different plant dynamics. Thus we have likely have an internal kinematic plan which we then adjust our controller model to ensure that plan is enacted, regardless of the differences in the plant.

With the opposite steer bicycle, there is a relatively simple change in the dynamics, i..e flipping the sign of relationship of applied force to rotational motion. But what would the kinematic plan be? Do we watch which direction the fork/wheel rotates about the steering axis? Seems like we more likely pay attention to the proprioception of our arms, but different length handle bars and front end geometry would mean we have to change our kinematic plan for our arms. The steer angle of the fork/wheel has to follow a kinematic plan to ensure that the vehicle stays upright. It would be interesting to test the rider's ability to control the vehicle without being able to see the fork/wheel position. Can they still adapt?



--
You received this message because you are subscribed to the Google Groups "Single Track Vehicle Dynamics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to stvdy+un...@googlegroups.com.
To post to this group, send email to st...@googlegroups.com.
Visit this group at http://groups.google.com/group/stvdy.
For more options, visit https://groups.google.com/d/optout.

andy ruina

unread,
May 6, 2015, 5:09:11 PM5/6/15
to st...@googlegroups.com
Slight correction:  Dan Ferris, friend of Andy, came accross this relevant paper. 
Andy just passed it on.




-Andy Ruina,   ru...@cornell.edu,  http://ruina.tam.cornell.edu
 cell:         607 327-0013,       Skype: andyruina


Anthony Doyle

unread,
May 10, 2015, 4:57:30 AM5/10/15
to st...@googlegroups.com

The question asked was what was the controlling variable when a human rides a bicycle? In the early eighties I researched into this question for my Phd.  I am not saying that what I found is universal because I only used two subjects but for those two the following applied.  First both were blindfold throughout the trial.  I subsequently found that a dozen competent riders chosen at random had no difficulty riding on an open space without hazards with their eyes closed though some of them needed quite a bit of persuading to try because they believed it to be impossible.  From this I concluded that although vision can play a part in bike control it is not necessary.  Both trial members had little difficulty riding an approximately straight course and none at all in keeping control at a slow speed.  At no time did either subject put a foot down or lose control during a trial run.  The bike was modified to remove all stability, that is the steering shaft was vertical and there was a counter rotating front wheel to remove its gyroscopic effect.  A highly sensitive accelerometer was mounted in the roll plane and this and the movement of the steering shaft were recorded.  The subjects brief was simply to ride slowly without falling off.  They were not required to ride in a straight line but they did. An analysis of the records showed that the correcting movement of the handle-bar was initiated when the roll acceleration reached a fairly constant threshold.  Because of the physical properties of man and bike the roll acceleration keeps on increasing with angle of lean.  The human vestibular system is very sensitive to accelerations but more or less blind to velocity as may be observed when travelling in a lift.  This means that the human control system cannot be modified by adding in some velocity.  Any attempt to keep moving the handle-bar till the acceleration stops leads to dramatic over-control as can be observed when watching someone trying to ride a bike for the very first time.  The strategy that is adopted, one that can be found in the control of other extremely unstable systems such as helicopters, is to make a more or less standard ‘spike’ of acceleration, a quick twist of the bar in the relevant direction followed swiftly with a return to its previous position.  This action injects a pulse of energy into the control system which either reduces or reverses the roll velocity. In most cases it reverses the roll velocity and the bike goes over the vertical and starts to fall in the opposite direction so that it proceeds in a series of oscillations in the vertical, something that could also be seen very clearly by examining the tyre tracks. The earlier the correction is applied the less it needs to be but it looks as though the pulse itself remains more or less standard. When more energy is required a second or third pulse is applied in quick succession.  All this takes place at time intervals that are well below the human conscious threshold which is why you can’t learn to ride a bike by thinking about it.

Riding a bicycle very slowly is difficult because the automatic stability effect more or less disappears and the rider has to provide all the control. When there is strong front wheel stability, control is quite different. For example, for normal road control on a motor-bike, where the offset steering axis and high gyroscopic forces produce a very powerful correcting response to both acceleration and velocity in the rolling plane, the rider simply alters the ‘zero setting’ of the system by pushing the handle-bar on the side to which he desires to turn. The immediate response is a roll in the direction of push which then provokes a correcting response from the automatic stability.  The rider keeps the push while allowing the bar to adjust is position under the influence of the three torque forces, fork geometry, gyroscopic precession and his push.  In doing this he is applying a torque that is independent of the bar’s position. The stronger the push the steeper the resulting turn.

For all the details google A J R Doyle Bicycle Riding; the theses is on the net.

andy ruina

unread,
May 10, 2015, 11:24:18 AM5/10/15
to st...@googlegroups.com, Anthony Doyle
Intro:

For those who do not know Tony, author of the email below, he is a thoughtful careful
scientist.  It would be reasonable to put high weight on his thoughts and observations,
so I think.





-Andy Ruina,   ru...@cornell.edu,  http://ruina.tam.cornell.edu
 cell:         607 327-0013,       Skype: andyruina



andy ruina

unread,
May 10, 2015, 11:55:31 PM5/10/15
to st...@googlegroups.com
Tony:

Do you think that roll-acceleration is basically the only sensing of level by
blind-folded humans?




-Andy Ruina,   ru...@cornell.edu,  http://ruina.tam.cornell.edu
 cell:         607 327-0013,       Skype: andyruina



On May 10, 2015, at 4:57 AM, Anthony Doyle <yodel...@gmail.com> wrote:

Damian Harty

unread,
May 13, 2015, 3:59:42 PM5/13/15
to st...@googlegroups.com
What does the aviation industry have to say on the matter? As I recall, people fell out of the sky willy-nilly when they flew into clouds because they couldn't detect a certain motion of the plane. It's why artificial horizons were invented. This makes me think our absolute attitude sensing is so poor as to be broadly useless for control input, even in the presence of gravity.

Jodi Kooijman

unread,
May 14, 2015, 10:48:30 AM5/14/15
to stvdy
I had students looking into which inputs we require to be able to ride a bicycle. In their search they talked to someone at the University or Maastricht (I've forgotten who he was exactly). The conclusion from experiments performed at the university Maastricht was: 

1) people with a defective or missing vestibular system can ride a bicycle - no problem. 
2) they however can't ride a bicycle with eyes closed. 
3) furthermore if they encounter a bump they have to stop immediately (Apparently your eye muscles are some how controlled by the vestibular system - so if you go over a bump without a working vestibular system your eyes pretty much bounce/rotate up and down in the socket.) 
4) they can't turn their heads to see if traffic is coming from behind (pretty essential in normal riding) for the same reason as 3. 


andy ruina

unread,
May 15, 2015, 7:47:57 AM5/15/15
to st...@googlegroups.com
I tend to agree with what Damian wrote below, as likely.  My question should have been
“Do you think that inner-ear roll-acceleration is basically the only thing
sensing deviation from steady straight, or steady-circular-path, riding?"



-Andy Ruina,   ru...@cornell.edu,  http://ruina.tam.cornell.edu
 cell:         607 327-0013,       Skype: andyruina



chester r. kyle

unread,
May 15, 2015, 2:44:24 PM5/15/15
to st...@googlegroups.com
This study group was started by Jason Moore in answer to my question about whether a blind person could ride a bicycle using the senses we have of location, acceleration, force etc.  Turns out they can, Jason located one who could ride a bike in a parking lot with some one following to give him instructions about directions.

Chet Kyle



Reply all
Reply to author
Forward
0 new messages