Gyro for controlled Wheelies

Willow

10 kW
Joined
Sep 6, 2011
Messages
814
Location
Tasmania
...Has anyone thought about using a gyro to enable a bike to maintain a wheelie - without taking it too far so that you would flip it. Give it full throttle, and your away.

It should be achievable through modifying the throttle signal... Segways effectively do it.

Anyone have suggestions??
 
yes. you are not the first one. :) me and several other mentioned that in various threads already. even justin may thing about this as an add on the his CA and controllers. implementations based on a simple arduino and gyro sensor as used in the quadcopters would be really easy to make. and quite cheap. an arduino one with sensor can be bought for $15 shipped.
 
BMW does it on their 200 HP motorcycles. They also have sensors for braking, lean angle, and traction control that affect the suspension rates. Such sensors would be great for front wheel drive on ebikes.
 
gogo said:
BMW does it on their 200 HP motorcycles. They also have sensors for braking, lean angle, and traction control that affect the suspension rates. Such sensors would be great for front wheel drive on ebikes.
such an anti skid implementation would be quite simple: a speed sensor on the front wheel, one on the rear wheel, and if the both differ too much you slow down the driven wheel. abs for a bike is not that simple at least not w/o adding additional hardware.
 
Get it working and reasonably priced and I'll buy some from you Willow.

BTW, Justin has said it would only require an inclinometer and a bit of programming for a CA3 to pull it off. It would be anti-flip as well as wheelie control.
 
not something I would do myself - but I may know some people who might be able to help...

All my (high power) bikes will be using Adaptto's - so no more CA... The device then, for me, must be stand alone - unless the Adaptto can manage it.

Though, a plug and play 'wheelie chip' for any electric bike..... would be wheelie good.
 
I think stopping the bike from flipping over backwards is only half of the problem for learning controlled wheel stands. The bike can also tip to one side causing the panic button to be hit. I guess the faster you are going the less this tipping to the side is an issue, but for an ebike and someone wanting the help of electronics to keep you in a controled wheel stand I think it is still going to take a skilled rider to pull this off . I think what stops most of us from doing a decent wheelie is fear (for good reason) and too slow reactions. I would love to try one of those fancy Audi bikes however. They must have very smooth thottle control which most of the home build bikes seem to stuggle with.

Also I think with the segway two wheelers the rider is pretty much standing on the axle so it would be much easier to control the tipping point. Maybe a good way to get comfortable with wheel stands is to hang foot pegs off the rear axle like bmx style and ride these foot pegs while practising wheelies to get comfortable with the tipping zone.
 
So essentially it should be a ardu box that plugs right in line of the throttle. It would then monitor gyro sensor and from set limits when bike is on both wheels till the safe point on one wheel it would pass through throttle signal and after that start to work the magic. You would need to tune it on every different system bike i guess.
 
there could also be an issue with NON DD hubs. the segway can accelerate and decelerate the motor. without the possibility to brake the wheel we are loosing one mechanism to control the wheelie. eg: cutting the throttle is a good way to stop the bike from flipping, but if the rider leans backward during the wheelie, cutting the throttle may not be enough to stop the flip.
 
I remember reading about a dirt bike wheelie record attempt in the 80's (?)... They rigged up some sort of electric motor to the front wheel to keep it spinning while it was in the air, I guess for stability?? I think they were trying to wheelie for quite a fair distance.
 
gm6046 said:
I remember reading about a dirt bike wheelie record attempt in the 80's (?)... They rigged up some sort of electric motor to the front wheel to keep it spinning while it was in the air, I guess for stability?? I think they were trying to wheelie for quite a fair distance.

The two greatest wheelie exponents in the 70s and 80s were Doug "Wheelie King" Domokos in the US, and Dave Taylor in the UK. The latter is famous for wheelying the entire Isle of Man TT course - that is 37.73 miles.

I believe both used a small motor to keep the front wheel spinning while it was airborne. Friction drive ! :wink:

http://www.motorcyclemuseum.org/halloffame/detail.aspx?RacerID=161
 
This can be done, but you have to understand physics and more importantly control theory. Without knowing control theory basics, you will not be able to make the device stable. I've been working on learning control theory now for about 2 months for a different project, but learning it made me realize when I thought about adding this feature to my throttle tamer I probably would have failed miserably. It's not as easy as it sounds, but it's not difficult once you understand control theory. You will need to have a very solid math background so brush up on your calculus and make sure you have a really good understanding of PID control.
 
ksithumper said:
gm6046 said:
I remember reading about a dirt bike wheelie record attempt in the 80's (?)... They rigged up some sort of electric motor to the front wheel to keep it spinning while it was in the air, I guess for stability?? I think they were trying to wheelie for quite a fair distance.

The two greatest wheelie exponents in the 70s and 80s were Doug "Wheelie King" Domokos in the US, and Dave Taylor in the UK. The latter is famous for wheelying the entire Isle of Man TT course - that is 37.73 miles.

I believe both used a small motor to keep the front wheel spinning while it was airborne. Friction drive ! :wink:

http://www.motorcyclemuseum.org/halloffame/detail.aspx?RacerID=161

Interesting...so the gyroscopic effect of the front wheel must be a significant factor in wheelie control.
 
zombiess said:
This can be done, but you have to understand physics and more importantly control theory. Without knowing control theory basics, you will not be able to make the device stable. I've been working on learning control theory now for about 2 months for a different project, but learning it made me realize when I thought about adding this feature to my throttle tamer I probably would have failed miserably. It's not as easy as it sounds, but it's not difficult once you understand control theory. You will need to have a very solid math background so brush up on your calculus and make sure you have a really good understanding of PID control.

I for one have to rely on you guys. The only control theory I want to understand is being a qualified enough controller user to make good buying decisions and optimize settings, along with making good requests for features from you experts.
 
John in CR said:
zombiess said:
This can be done, but you have to understand physics and more importantly control theory. Without knowing control theory basics, you will not be able to make the device stable. I've been working on learning control theory now for about 2 months for a different project, but learning it made me realize when I thought about adding this feature to my throttle tamer I probably would have failed miserably. It's not as easy as it sounds, but it's not difficult once you understand control theory. You will need to have a very solid math background so brush up on your calculus and make sure you have a really good understanding of PID control.

I for one have to rely on you guys. The only control theory I want to understand is being a qualified enough controller user to make good buying decisions and optimize settings, along with making good requests for features from you experts.

Amen to that.
 
My thought was that it is less complicated. Poster wrote that this device should limit wheelies not to flip over and not the way segway or lit motors balancing vehicle work which would be a lot more to achieve. As I understand it should limit throttle input based on gyro reading
 
Punx0r said:
Got to be easier than the classic inverted pendulum control theory exercise?

Anti flip can be done using a bang bang type control, but it won't be smooth and you won't be riding out a wheelie. Even dialing in an anti flip could get tricky because the device should really monitor the rate of the front end rising. If it happens quick enough it could put enough energy into the system to still flip.

A simple anti flip setup could apply regen to compensate without knowing the rate of change, it could say if your at this angle, apply regen.

To really get wheelie control, you need good throttle control. Speed only control that most controllers use makes it more difficult, but not impossible. The other gotcha is what happens if the bike is on non level ground, this should be detected and taken into account because it takes less effort to flip when starting off on an uphill.

I seriously considered adding wheelie anti flip into the throttle tamer but lacked the know how until recently. It has been on my possible project list for a while and I even purchased some sensors, but I have been wrapped up in some larger projects that others are counting on me for.

I would probably start off withan inverted pendulum setup and work from there. The goal would be to maintain a certain difference angle from the bicycles "level" reading as long as enough throttle was commanded (a run away wheelie would suck). Any time you interface to a critical system such as a throttle you have to think through all scenarios, failures and how to handle them.
 
If you're using one of the 3D acceleration sensors like most phones/etc have in them nowadays (and are included in things like the old STM Circle and it's successor MCU dev platforms), it should be easy enough to detect relative angles and accelerations, even with only one sensor inside a single "box" that just plugs into a bike's existing system to do all the wheelie control. If you can use two, one at the front and one at the back, it might be even easier as you can simply compare separate data sources.
 
as i said using a 3 axis gyro and accelator meter like the quadcopters use it would be the solutions with less parts. you may just let the system learn what is level and then you have the angle relative to that all the time. you get acceleration of all 3 axes from the chip which has to be put into realtion to level.
 
izeman said:
as i said using a 3 axis gyro and accelator meter like the quadcopters use it would be the solutions with less parts. you may just let the system learn what is level and then you have the angle relative to that all the time. you get acceleration of all 3 axes from the chip which has to be put into realtion to level.

Yes, a 3 axis accelerometer is a great way to go and I even purchased some for the specific purpose of interfacing to the throttle. Getting the parts to accomplish the task are easy and cheap, coming up with the code, the math and the controls to provide a stable system are much more difficult.

I spent a solid 2 weeks learning how to identify a system, model it and create a PID loop. I'm only about ankle deep in control theory. I have an appreciation for it and have just started to understand how useful the Laplace transform is and how linear and non linear systems behave. I've finally connected state space with the poles and zeros involved in the matrix math.

As an exercise, it would be good to come up with a feature list of what the wheelie control needs to do, what situations will cause it issues and possible solutions to those issues. Then write out the sentences on how the controller would handle them, this will become the pseudo code.

The main reason I purchased the 3 axis accelerometers is I wanted to make an acceleration based throttle for the xie chang controllers to make it's response more predictable.

So to make an anti wheelie controller, how do we make the throttle control more predictable? My own personal thought was to translate it to acceleration in G force in the forward direction while using the accelerometers orientation as a reference to "level" ground. This way I'd be working in all G force units to make the math units the same.
 
Good points.

My understanding of control theory is very basic. If you have a copy of Matlab then there's a program in there called Simulink which allows you to model control systems using drag-and-drop block diagrams, with the output as a scope trace. Very simple to use and play around with.

Is the idea behind using a three axis accelerometer that you can subtract acceleration in the forward axis from the vertical (to differentiate between the bike's angle and it's forward acceleration)?
 
You need gyro and accelerator to know a) your position in all 3 axes and b) to know the degree of change of all those three.
Knowing only if the bike is level to ground means nothing. You could stand on level ground or you could go downhill wheeling and the result for the microcontroller would look the same. But if you add the acceleration around the right axis then you know that you may be close to flipping over.
 
Back
Top