Motorcycle/Rider simulation. Problems controlling oscillations.

[BigShox]

Hello, My first post and it's going to be a question =)
I hope I can get some replies.
Here we go :- I'm constructing a simulation of a motorcycle with rider and I have encountered a problem with oscillation (lack of stability) that I just can't find a solution for. Basically the bike will run in a straight line fine, and it will turn and lean into a corner ok (I'm currently working my model with two masses - One for the bike and one for the rider, I'm balancing and leaning the bike by moving the mass of the rider around). But it then gets into a cycle of oscillation that eventually leads to the bike falling over or getting into even worse conditions of pure instability. The problem centers around the model slightly overshooting it's desired lean so the mass of the rider constantly shifts from side to side to compensate thus leading to the oscillations. I've been trawling the net, books and other sources and have reached a brick wall. There must be a solution out there somewhere =)
Hence my post here, to get some other minds on the problem and hopefully a post or two that will get me off in the right direction (so to speak).
btw - great forum you have here..

 

[berkeman]

Fun topic, and welcome to the PF. What rake and trail are you running on your bike model? Have you looked at the sensitivity of these oscillations to rake, trail and ride height front/rear? Oscillations like those you are seeing are very real in real world design of motorcycles, and you should be able to tune them out with variations in ride geometry (that's what the industry has been doing for a number of years now).

Also, try adding a steering damper function to the front steering bearing arrangement. It's common for race bikes to use them (I have one on my main racetrack sportbike), and it will give you an added variable to tune. You can even give your damper both high-speed and low-speed damping characteristics like my Scotts steering damper.

There's nothing like skipping the front wheel of your sportbike while leaned over at a buck to make you appreciate a good steering damper! Let us know what you find.

 

[Homer Simpson]

Is this a computer model or real one? What is the control system for the lean? If it is a P-I-D type of control system on a real model it would be bound to overshoot, and over correct. Do the bars rotate unassisted or are they controlled? Look into the concept of 'push steering' and 'speed wobble'.

 

[Claude Bile]

The problem does not appear to be in the model, but in the programming, since the simulation is making the 'rider' lean more than the model says it should. I would take a closer look as to why the program leads to the incorrect result. If you remove the overcorrection, you remove the drift into chaos.

It may be that you cannot remove the overcorrection completely, in that case in may be practical to add a damping term to prevent the positive feedback loop that appears to be driving your model toward chaos. You should attempt to justify the damping term physically to avoid making it look like a fudge factor.

Claude.

 

[krab]

I suspect there's nothing incorrect in the model as far as it goes, but that you don't have enough damping in the description. How are the two masses linked? In real life, the body of course damps this linkage. So try it with some damping between the masses.

 

[rcgldr]

Duplicate post, no response after posting the first time.

 

[rcgldr]

There's a basic flaw in your model. A rider leaning to steer a motorcycle works, but only because of self correting steering geometry, and at high speeds (100+mph) gyroscopic forces resist change in lean angle, slowing down the recovery repsonse to the point that it's virtually gone, and trying to lean with your body won't work.

The reason you see most modern motorcyle racers hang off the bike is two basic reasons. One is that it's a bit more stable if the tires slide a bit. Two is that if the motorcycle slides out from under you, you have less distance to fall. You'll should also note that the riders hang-off before leaning the motorcyle, using counter-steering to keep the bike vertical while shifting their body weight and keeping it vertical after hanging off until they enter the turn. Back in the old days, when crashing a few times per season wasn't accepted as part of the job of motorcycle racer, they didn't push the bikes to the limits as hard and didn't lean with their bodies at all. It was all counter steering input controlled.

Motorcycles lean because the front wheel is turned away from the direction of lean. This happens due either directly to rider input on the handle bars, or because if the rider leans one way, the motorcycle initially leans the other way, and the self correcting steering geometry causes the wheel to steer outwards, and now since the center of mass is off to the lean side, the self correction geometry won't completely straighten up the bike.

The self-correcting steering geometry is due to the fact that the contact point is behind where the pivot axis intercepts the ground. Lean the bike over even while standing still, and the front tire "falls" into the direction of the lean (I suggest you do this experiment with a bicycle, which is a lot lighter). Also, if you steer left, the contact patch translates right, and vice versa. You can test this yourself, while standing still on a bicycle or morcycle, steer left, and you note that the front end moves left, and vice versa. This is because the contact patch translated in the opposite direction, and since it doesn't slide the front end of the bike moves in the same direction that you steer.

There are two terms used in steering geomertry. Trail is the distance from where the pivot axis intercepts the ground back to the center of the contact patch. Rake is the amount of angle of the pivot axis. Both affect stability, and self correction. To reduce steering effort, usually the forks are placed in front of the pivot point, which reduces the trail, or on a 10 speed, the forks are just curved forwards at the bottom. Radio control motorcyles move the contact patch futher back. They work because they use a lot of stability, if you hold one up and steer it, you'll note that the front tire translates by a very large amount.

Note that the higher the speed, the slower the self-correction response. At high speeds, it takes almost as much countersteering torque to straighten up as it does to lean over, because the gyrscopic forces resist any change in lean angle. Since it's the self correction response that allows a bike to be leaned by the rider leaning, rider leaning doesn't work at high speeds. There is a transition in the type of stablity versus the speed. At low speeds, you have self-correcting stability, ease up on the handlebars, and the motorcyle straightens up on it's own. At high speeds you have anti-lean stability, the gyroscopic forces resist any change in lean. So as speed increases, there's a gradual transition from self-correction stability to lean stability.

If you own a motorcycle, try slight movment of the handlebars back and forth to lean the bike side to side while essentially traveling forwards. At 40mph there's little resistance in the bars. At 70mph you can feel a significant resistance, and if you can find a place you can try this at 100+mph, the preceived effect is that you apply a lot of torque the the handlebars, but that they don't seem to move at all. Eventually you associate counter-steering pressure with lean (as opposed to actual movement of the bars). This is the goal, since it works at all speeds on conventional motorcycles.

The exception is a bonneville type bike, which has a very low center of mass; there will be a speed range, usually between 100 and 200mph where the bike control inputs are not reversed, and outside of this range, the typically reversed counter steering inputs work, and apparenlty the transitions are not that difficult to deal with, since you're just trying to avoid any leaning other than to keep following a straight line.

If a steering geometry is prone to over correction induced oscillations, then a damper is used to prevent this.

If you've ever watched a motorcyle race over a rough course, like Isle of Mann, you often see the bikes get disturbed by ireggularties in the pavement causing some pretty serious wobbles, but that the bikes self correct and the wobble goes away. You can test for wobble by letting go of the handle bars and banging one of the bars with your hand to induce a wobble. I've only owned one street bike (Suzuki GS1150ES), where the wobble would get worse (re-grabbing the handlebars stops the wobble so it's not as risky as it sounds, just don't panic). This bike would have needed a steering damper for high speed racing. The mass of the riders arms is usually enough to dampen out wobbles on most motorcycles, but a steering damper is always good for high speeds, higher speeds usually result in more severe wobble response, depending on overall geometry.

In the old days, excessive flex in the framework, especially forks and swingarms, made some bikes (like the old Royal Enfield) unstable at high speeds, and these bikes would get into an unrecoverable wobble. In an unrecoverable wobble, it eventually gets to the point where the wheels strart hopping side to side and up and down, eventually spitting the rider off the bike.

 

[berkeman]

There's a basic flaw in your model. A rider leaning to steer a motorcycle works, but only because of self correting steering geometry, and at high speeds (100+mph) gyroscopic forces resist change in lean angle, slowing down the recovery repsonse to the point that it's virtually gone, and trying to lean with your body won't work.
The reason you see most modern motorcyle racers hang off the bike is two basic reasons. One is that it's a bit more stable if the tires slide a bit. Two is that if the motorcycle slides out from under you, you have less distance to fall. You'll should also note that the riders hang-off before leaning the motorcyle, using counter-steering to keep the bike vertical while shifting their body weight and keeping it vertical after hanging off until they enter the turn.

Jeff has some interesting points, but he's wrong about the racetrack stuff. That's okay.

 

[rcgldr]

Jeff has some interesting points, but he's wrong about the racetrack stuff. That's okay.

Wrong how on the racetrack stuff?

I didn't get into details, like one of the reasons for increased stability when hanging off on the inside of a turning motorcycle is that it moves the contact patch futher away from the edge of a tire, which may improve grip a bit and/or allow for recovery if the tire slides a bit.

On street based motorcycles, you had to hang off to keep parts from scraping, but on GP class racing bikes, there's usually more than enough clearance to allow the bikes to lean over to maximum lateral force without hanging off.

Hanging off uncoordiantes the turn, the bike is not leaning as much, and this introduces some lateral force on the axis of the wheels. I'm not sure if there's benefit to this.

Obviously it's worse aerodynamic wise to be hanging off, so no one hangs off on really high speed turns, like going on to the banked oval at Daytona, the riders tuck in while still cornering (then again they're on the banked section by then, looking nearly horiztonal). Regarding the counter-steering effort, I can't remember the exact number but at 180+mph coming out of that banked corner at Daytona, it's take a very large amount of counter steering torque to get the bikes to straighten up again.

 

[berkeman]

Wrong how on the racetrack stuff?

Hanging off does a number of things, but mostly it's for ground clearance (yes, even on Rossi's GP bike), and to stay balanced on the bike with the centripital acceleration. On the racetrack, you typically only have about 20% of your body weight on the seat, and most of the rest of it down on the footpegs (the pivot point of the bike in pitch and yaw). When you're way over in a turn, you have a large part of your body weight on the inside peg, and you're pulling the bike over with your outside thigh and pushing it down with your inside foot. Your inside knee puck is just a feeler for ground clearance -- you don't put any weight on it while it's sliding on the pavement. You shift your body weight to the outside peg as you spin up at the corner exit...kind of like how you weight the outside peg on a dirtbike as you slide corners.

There is an ongoing debate among sportbikers whether countersteering or body steering is best, but I'm a big proponent of body steering. You don't steer the bike with the handlebars much at all when using body steering, and instead use peg weighting and body weight shifts (plus pulling in with the outside thigh) to lean the bike, and let the natural stability of the bike point the front tire mostly by itself. At my first racetrack school on a sportbike, I found out that I couldn't countersteer fast enough to make transitions into fast combination corners like Laguna Seca's Corkscrew, and that if I just relaxed and used my body weight shifts to steer, the bike took care of the rest.

So at least from my perspective, and based on about 1000 miles of racetrack riding, you hang off of a sportbike in corners for ground clearance, to turn the bike with body steering, and to keep your weight centered in the middle of the bike in the corner. If you start pushing the front or losing the rear in the corner, you are in the best place to deal with it if your weight is centered and neutral against the g-forces.

http://www.classrides.com/

http://www.starmotorcycle.com/

 

[rcgldr]

Note: total re-edit of this reply:

ongoing debate among sportbikers whether countersteering or body steering is best

A motorcycle is a unitrack vehicle; all changes in the roll axis are due to the tires being steered outwards from under the center of mass, plus some gyroscopic reaction at the front tire (which fortunately generates the same lean response as counter-steering).

What body steering really does is allow a rider to build up angular momentum in the roll axis by shifting his body inwards while counter-steering (to prevent the bike from leaning the wrong way), and just a bit later, use this built up momentum, resulting in a delayed, but faster rate of lean of the motorcycle. The overall rate of lean of the system (bike and rider) is always due to offset lines of forces, which is what counter steering induces, which in turn produce a torque force on the roll axis, but by shifting his weight inwards, a rider can delay the lean response of the motorcycle component of this rider/bike system, and then transtion into a mode where the rider stops shifting inwards and the bike's delayed rate of lean is sped up.

So the overall rate of lean is strictly due to counter-steering, but by shifting his weight inwards, a rider can apply counter steering forces sooner, and therefore, over a longer period of time, than if the rider didn't shift his weight inwards, and the result is more lean angle by the time you enter the corner.

weight on the inside peg

If a rider is hanging off, the line of force vertical to the bike/rider from the peg probably passes by the outside of the center of mass of the rider. What's keeping the rider from falling off the bike due to the inwards torque from hanging off are sideway forces from his hands and outside leg applied to the handlebar and sides of bike. These same forces are also applying a lean force to the bike, but any change in lean will always be due to an offset between the lines of reactive and actual forces on the bike/rider, which is controlled through counter steering inputs.

 

[rcgldr]

ongoing debate among sportbikers whether countersteering or body steering is best

From my previous post, it's clear that body shifting (steering) if timed properly, can allow a rider start counter-steering sooner, which allows a rider to achieve more lean angle sooner than without the body shifting. So the real point here is where on a track are you pressed for time to achieve a lean angle? Fast s-turns transtions at near constant speed would be such a situation. But most turns on tracks have a lot of setup time, especially on typical, brake, apex, accelerate type turns.

One point I missed in my previous post is that a rider can also raise his center of mass before shifting sideways, which would increase the amount of angular moment, then lower it before leaning the bike, decreasing the angular moment, allowing the built up angular momentum to have more effect (like a diver tucking while flipping in a dive).

The other point is no-one truly "body steers", when they shift sideways, they are counter-steering to prevent the bike from leaning the wrong way.

True "body-steering" would be the equivalent of riding with your hands off the handle bars, or, a free turning front end with fixed handlebars.

 

[BigShox]

Thanks for all the feedback!
Homer - it's a computer model. It's actually based on a quite successful (as in it works ) model we built for simulation of 4 wheeled vehicles and we're working more into the system to incorporate 2 wheeled vehicles.
The program so far works like this:-
There is a core physics and collision system that deals with gravity, friction, restitution etc (it's basically an impulse based physics simulator). On top of this there is a dynamics system that deals with the mechanical aspects of a vehicle, from the contact patch through the tyres and suspension system, driveshaft, gearing, engine, steering etc. An AI system (or user) then controls the vehicle through throttle, brake and steering input. The behavior of the vehicle is a natural outcome of the interaction between the physics and dynamics models.
All this was fine for our 'cars' but causing us problems for the bike.
That said there's been some really useful info here, thanks guys, and it's opened up a new direction in thought.
How we were doing it - The bike was fine and very happy to go in a straight line, to corner we were shifting the weight of the rider and allowing the wheel to turn, causing the bike to 'fall' into the corner. The angular acceleration involved was leading to a constant overshoot of a 'stable state', much worse when coming out of the corner and leading to the bike crashing not long after trying to straighten up (even with damping).
We're now trying an approach that is slightly different by starting the cornering action with a small steering input and then moving the 'rider' to stop the bike falling 'out' of the corner. This has greatly reduced the oscillations and at slow speeds there is no noticeable oscillation at all, however currently the model does not want to turn very well at higher speeds and is tending to crash (by tumbling away from the corner). It's a very interesting problem (even though it is frustrating).

 

[berkeman]

The other point is no-one truly "body steers", when they shift sideways, they are counter-steering to prevent the bike from leaning the wrong way.
True "body-steering" would be the equivalent of riding with your hands off the handle bars, or, a free turning front end with fixed handlebars.

This is why I phrased my mention of body steering the way I did. People who haven't learned body steering don't realize that you don't need to countersteer to turn a sportbike. I often ride with zero force on the bars, and only add a roll yank in for faster transitions. I try to never use a yaw push or pull on the handgrips to turn in. The bike turns because it's leaned, and you can lean a bike no problem with your hands off the bars.

I used to use countersteering, but the racetrack schools were real eye openers. The biggest reason that I've stopped using countersteering and focus exclusively on body steering is that in emergency situations, your arms tend to tense up, and if you have to use countersteering to turn, you have much less control for evasive manuvers. With body steering, your arms stay relaxed and loose, and you basically just jump side to side to get around the problem. If anybody here in the PF already rides the street and would like to have an incredibly fun and valuable day, I highly recommend taking a racetrack riding school, especially the CLASS and STAR schools that I posted links to earlier in this thread. Ride safe!

 

[berkeman]

We're now trying an approach that is slightly different by starting the cornering action with a small steering input and then moving the 'rider' to stop the bike falling 'out' of the corner. This has greatly reduced the oscillations and at slow speeds there is no noticeable oscillation at all, however currently the model does not want to turn very well at higher speeds and is tending to crash (by tumbling away from the corner). It's a very interesting problem (even though it is frustrating).

Hi Bigshox, cool stuff. Two follow-up things. First, I don't think you mentioned what bike geometry you are using. Are you using a standard sportbike geometry, with rake, trail, wheelbase, ride height, etc. matching a real sportbike? Modern sportbikes have basically all moved toward a set of numbers that have been proven on the racetrack as near-optimal for stability and agility.

Second, even though I'm a big proponent of using only body steering, are you guys in your project familiar with the concept of countersteering? It's the basic steering technique taught in the beginning riding schools like the Motorcycle Safety Foundation classes, and it works fine for turning a motorcycle as long as your arms don't tense up. Using countersteering, you push the right handgrip forward (turning the front wheel left in yaw) to initiate a right turn. Making the front wheel point briefly to the left moves the front end left and out from under the center of mass of the bike and rider, which tips the bike to the right to start the right turn. As you reach the desired lean angle, you ease off pushing on the handgrip and stabilize at whatever lean angle is appropriate for the turn.

If you don't have active steering inputs going into the handgrips yet in your model, you might try adding that in. A steering damper is still a good idea though.

 

[rcgldr]

How we were doing it - The bike was fine and very happy to go in a straight line, to corner we were shifting the weight of the rider and allowing the wheel to turn, causing the bike to 'fall' into the corner. The angular acceleration involved was leading to a constant overshoot of a 'stable state', much worse when coming out of the corner and leading to the bike crashing not long after trying to straighten up (even with damping).

This could be due to too much correction factor, and/or not enough gyroscopic reactions. The pivot axis for a typical motorcyle is around 25 degrees or more from vertical (25 degrees of "rake"), to reduce the perpendicular (to fork centerline) force the forks experience when braking and/or going over bumps. This results in too much correction factor, so the forks are mounted forwards of the pivot axis, to reduce the "trail" which is the distance from where the pivot axis intercepts the ground back to the contact patch.

As long as there's some amount of trail, the front tire translates sideways in the opposite direction that it's is turned. Pull the front end of a stationary bike to the left and the front wheel turns left. Lean a stationary bike to the left and the front wheel turns left because gravity does the pulling for you.

Hold a bicycle by the rear seat and tilt it sideways and the front tire will fall into the turn. Note that you can only lean the bicycle a small amount before the front wheel turns and translates so much that you can't straighten the front wheel by leaning the bike to the other side. This may be why your model get's stuck in an overcorrecting response mode.

In real life, the self correcting steering geometry would only work well within a small range of speeds. However gyrocopic forces reduce the self correction repsonse as speeds increase, and the motorcycle transitions from a tendency to straighten up towards a tendency to maintain a constant lean angle, so this expands the high end of the speed range. Be sure that your model simulates gyroscopic forces.

There is no simple fix to expand the low end of the speed range. Increasing the trail reduces the entire speed range, but if it's lowered too much, then gyroscopic forces won't prevent overcorrection at higher speeds. Below the low end of the speed range, there isn't enough self-correction factor. The rider is forced to used counter-steering or his legs to prevent a fall. If you want to learn how to drive slow, learn to drive a 10 speed bicyle very slow until you can balance it while completely stopped (the sideways translation of the wheel let's you balance by counter-steering even when you're stopped). Trials motorcycle riders can hold a position while stopped, but they "cheat", in addition to counter steering, they use one leg to generate rotational impulses, similar to a high wire walker.

We're now trying an approach that is slightly different by starting the cornering action with a small steering input and then moving the 'rider' to stop the bike falling 'out' of the corner. This has greatly reduced the oscillations and at slow speeds there is no noticeable oscillation at all, however currently the model does not want to turn very well at higher speeds and is tending to crash (by tumbling away from the corner). It's a very interesting problem (even though it is frustrating).

The rider doesn't need to move at all. The model would work beter if you just smartened up the rider so that all lean angle changes were controlled by counter steering.

Radio control motorcycles work without any rider shifting. One thing that some models do is cheat. There are wheels on the side of the model that limit the lean. At slow speeds, where self-correction doesn't work, the models just rest against the side wheels and hold the lean until you speed up enough for self correction to work. The steering geometry is also bizaare in that the front wheel translates more side to side than it pivots (probably the forks are angled backwards realtive to the pivot axis to get this effect, I'll have to inspect a model to be sure).

However, a "smart" rc model could use a crystal based "gyro" that senses rotation about an axis, and use it for corrective feedback control on steering inputs. There are smart "heading hold" gyros for rc helicopter that feed in yaw control inputs to hold a heading or to hold a rate of heading change inspite of side loads on the tail rotor, allowing a rc heli to make smoothly yaw 360 degrees while flying in a straight line even at high speeds. A similar "lean hold" gyro could do the same thing by counter-steering the model in order to initiate and hold a lean as requested by the transmitter.

If you're simulation incorporated the equivlent of a lean hold gyro, you're stability problems would be solved. Note that these gyros have a sensitivity and dampening adjustments, so there is some "servo" logic internally to prevent oscillations.

 

[rcgldr]

you don't need to countersteer to turn a sportbike.

This simply isn't true. It's a uni-track vehicle, like balancing a broom stick, you can't just move it left without at first moving the bottom a bit to the right to initiate a lean.

The contact patches of a motorcyle provide virtually no resistance to roll axis changes. The rider can't translate the center of mass sideways without counter-steering input. If the rider shifts his center of mass left, the the motorcycle's center of mass just shifts right in response, and the center of mass of combined rider and motorcycle doesn't translate at all no matter what the rider does (unless the rider jumps off the bike). The rider can't shift the center of mass forwards or backwards very much because the only resistance to this is the angular momentum in the wheels and tires. The only direction the rider can translate the overall center of mass is vertically (by raising or lowering his own center of mass).

The only reason that body leaning works is because the self correcting steering geometry is sensitive to the lean of the motorcycle, and not of the combined rider/motorcycle system. So when a rider shifts his center of mass to one side, the bike leans to the other side and the self correction response straightens up the bike. Now because the center of mass is offset to one side, the motorcycle continues to lean until the self correction process compensates enough for the translated center of mass.

As I've just posted, self correction only works within a certain speed range. It will never work at speeds slower than the operating speed range for self correction. At high speed, the rate of self correction is greatly reduced to the point that it's almost non-existant. The gyroscopic forces of the front wheel/tire respond by turning the wheel inwards in response to a lean which is self-correcting, but the rear wheel/tire gyroscopic forces act as a huge lean dampener.

I don't understand why you think letting the bike counter steer for you by leaning the bike the wrong way is better than just counter steering directly. The repsonse rate is slower if you let the bike counter steer for you, and your stated goal was to increase the rate of lean angle change.

 

[berkeman]

This simply isn't true. It's a uni-track vehicle, like balancing a broom stick, you can't just move it left without at first moving the bottom a bit to the right to initiate a lean.
The contact patch of a motorcyle provides virtually no resistance to roll axis changed. The rider can't directly translate the center of mass by leaning. If the rider shifts his center of mass left, the the motorcycle's center of mass is shifted right, and the center of mass of the combined bike and motorcycle doesn't change.

Sure it is. Just try this -- as you're riding along, pull with your left heel towards your right. This pulls the bottom of the bike to the right, which moves the contact patches out from under the center of mass, and the bike leans left. Yes, the bars then turn left as part of the lean, and stabilize at the correct yaw angle to track around the arc of the turn that corresponds to the lean angle and centripital acceleration. To speed up the turn-in, pull up on the right handgrip and push down on the left handgrip, which rolls the bike to the left. You don't need to turn the wheel at all to initiate a lean and turn.
Just think about it -- if you need to use countersteering to turn, then how come we can ride wheelies forever, including around turns? The front wheel is out of the picture at that point, isn't it?

I don't understand why you think letting the bike counter steer for you by leaning the bike the wrong way is better than just counter steering directly.

The bike is not countersteering itself. The phrase countersteering comes from the non-intuitive active turning of the front wheel the opposite way that you want to go. With body steering, you lean the bike with weight shifts, and the front wheel then points down into the turn and settles at the correct turn angle to match the lean angle.
Like I said, the main two reasons are that it's much more natural and easier to steer a sportbike on a racetrack with body steering, and because relying on countersteering puts you at risk of a crash in emergency situations where your arms tense up as a natural reaction to the danger. Take the Corkscrew at Laguna Seca, for example. At my first racetrack class, I tried over and over to countersteer down the best line through the Corkscrew, and I just could not hit it. I was pretty darned good and practiced at countersteering, but the combination of the speed, the blind rise/bump entry, and fast right-left-right transition combination kept putting me left or right of the fall line going down the hill. It was very frustrating! But by about mid-day of the class, the repeated encouragement of the instructors to stop trying to force the bike around with countersteering and instead just use your body weight really started to connect with me. It is so much more natural to just jump through turn transitions like that, control where you want your body to go, and let the bike do all the steering (pointing of the front wheel) for you. What a revelation.

And one of the things that bothered me so much about countersteering was that part about tensing up in emergency situations, and how that compromised my ability to keep steering to get out of danger. Like when you get into a corner too hot, or you get part-way around a corner, and all of a sudden there's something in the road that forces you to decrease the radius of your planned path a lot, or when you have to brake hard and swerve all at the same time to avoid a fool in a car that just turned in front of you. With countersteering, you can start the manuver usually okay, but any further additions to the maneuver get really hard as you get the natural tension in your arms as you try to use your arm muscles to save your butt. But when you practice body steering all the time, and consciously avoid using countersteering, your arms are relaxed all the time, including in emergency situations, and you basically have open relaxed hands working the controls. I honestly can say that switching from countersteering to body steering has saved me from several crashes on the street, including some pretty amazing evasive maneuvers.

Do you ride, Jeff? If so, have you considered taking one of the racetrack classes? They are definitely very worthwhile. Heck, about half of them say not to use body steering, and to focus on countersteering instead!

 

[rcgldr]

pull with your left heel towards your right. This pulls the bottom of the bike to the right, which moves the contact patches out from under the center of mass

The contact patches only translate sideways because the trail from the steering geometry responds to side forces at the front end by steering the front in the same direction as the side force. If the front wheel was not allowed to turn, then the contact patches wouldn't move.

Just think about it -- if you need to use countersteering to turn, then how come we can ride wheelies forever, including around turns? The front wheel is out of the picture at that point, isn't it?

Yes the front wheel is out of the picture, so this means that the rear tire is countersteering.

I think you're missing the point here on uni-track control; do a web search on the physics of unicycles and you find references to countersteering (plus yaw impulse or "reaction" steering). The same principle still applies: the contact patch can't be moved sideways (this would require sliding the tire), so it has to be steered sideways out from under the center of mass in order to initate a lean from a vertical position. Once leaned, then the offset forces from the pull of gravity on the center of mass and the pavement pushing up results in a leaning (roll axis) torque.

Getting back to doing wheelies, you steer by yawing the motorcycle. By moving mass that is in front or behind the contact patch, you yaw the motorcycle to countersteer. Curvature in the rear tire will results in a natural rate of turn versus lean, but you have to match this with the right speed.

In the case of a unicycles, the contact patch has some yaw resistance, epecially when stopped or moving very slow, so varitation in yaw impulses can be used to spin. The proper way to turn on a unicycle is to use variation in left / right foot pressure to countersteer.

Getting a bit off topic here, if interested, you can do a web search for robotic control systems for unicycles and/or motorcycles. (More interest in the unicycles, since the motorcycle case is simple enough that radio control models can be made that aren't very "smart").

Getting back on topic here, a motorcycle can be controlled without any weight shifting, relying on just counter-steering inputs to control lean.

what do you drive

I've never raced wheel to wheel, and only rarely got some track time, but I used to do wheelstands on a regular basis in my younger days, and I've owned motorbikes / motorcycles since I was 14 (you could get a permit at 14 in Texas back in the 60's), and I'm now 54 and own a hayabusa:

Some 0 to 100mph runs demonstrating the front wheel lifting at around 45mph (redline in 1st is 81mph). The goal here was acceleration, not wheelies, so I'm modulating the throttle to keep the front end in check. If I'm just having fun, I'll lift up the front end 2 feet or so, but I don't want to mess up the fork seals and or bearings so I don't do wheelstands anymore as you sometimes come down pretty hard (especially since it's been more than a decade that I did my last wheelstand).

http://jeffareid.net/real/busa.wmv

 

[berkeman]

Fun video. Uh, what's that 45mph sign doing there...

You really would enjoy taking a CLASS or STAR school day, Jeff. Definitely check out the links. Keeping the fast stuff on the racetrack has really helped me keep my street riding a lot safer. Well, except that now I ride motocross more (at 48 years young), and I'd rather spend a day on the MX track instead of on a sportbike.

Hey, BTW for the OP, have you looked at the "GhostRider" project, where they built an autonomous robot motorcycle to compete in the unmanned cross-country race competition recently? There's a very real world existence proof that you can make it stable...

www.ghostriderrobot.com[/URL]

 

[rcgldr]

I think part of the "panic issue" you referring to is due to a rider trying to force the motorcyle to follow a certain path. Since you can't do an uncoordianted turn for more than a fraction of a second, the only control you have over where a motorcycle goes is lean angle, and lean angle is most directly controlled through deliberate counter-steering.

Getting off topic, the same thing occurs other sports. In tennis and espically table tennis, you have to learn a proper stroke motion and learn to adjust this proper motion instead of trying to aim the ball which just doesn't work. The idea is similar, you have to learn to associate proper control inputs for where the ball will end up going.

Same thing with a motorcycle, you have to learn to associate lean angle with where motorcylce is going to end up. In the case of racing, you have to learn when and where to initiate turn in so that you end up taking the best line.

 

[rcgldr]

Fun video. Uh, what's that 45mph sign doing there...

Just decoration that day, the road was sort of shut down that day.

Keeping the fast stuff on the racetrack has really helped me keep my street riding a lot safer.

I'm more into the acceleration than speed, and I'm aware of the speed square relationship between braking distances versus speed, so I keep my speeds down. Last speeding ticket was back in 1984, althouh I do occasionally take advantage of opurtunities to do short speed runs (up to 100mph or so) when they show up, usually due to accidents leaving long stretches of unoccupied freeway before you get to any on ramps to worry about. The longest stretch I ever encoutnered was about 2 1/2 miles long, but unfortunately I was driving my wifes car instead of the bike that day. So almost all of the time, I just go with the flow, and enjoy the 0 to 75 or so runs on freeway on ramps. I live in So Califorinia, so the main concern with going 75mph is getting rear-ended, espeically with our daily high speed pursuits.

robotic control

That's impressive that they can control a dirt bike. Especially recovery after hitting objects. That sucker must be tough because it crashes a lot (look at the flip video). Now if they can get a robot to do MX, and triple jumps, I'll really be impressed. In a lab situtation, there are pogo stick robots than can do flips, these are pretty impressive, as it's only in the last decade or so that true dynamic control where there is a lot of off balance control inputs and outputs have been worked out. The gimbal control they use is a 2 axis gimbal (it has roll and pitch, but no yaw, since it's a motorcycle and not a unicycle (or helicopter), you don't need a yaw axis anyway (unless you want to do donuts)).

 

[rcgldr]

other vehicles I've ridden

Forgot to include this in my other post. I borrowed a friend unicycle and was able to ride it in about 5 minutes, since even before riding it I assume the main thing was to keep your weight in the seat and only use light pressure in the pedals. I was using the arm / impulse method for reaction steering. I didn't borrow it long enough to figure out that variance in pedal pressure is how to counter steer it.

As a teenager I remember riding a simple device that consisted of two 9 inch or so size wheels connnected by a bent up axis with 2 pedals. There was a similar contraption with 4 wheels, but the two wheeled one seemed more intersting.

I also remember when the handles of home built scooters broke off, they became the first skateboards. I went through the transition from steel wheels to composite made up from orange peels, to urethane. Never rode skates with wooden wheels though.

I've also ridden the large 3 to 4 foot in diameter wooden spools used for cables. It's only one axis of freedom, and the point of control, was the center of the spool, so the control inputs weren't that delicate like they would be for trying to ride a large empty ball (if it had a lot of inernal mass, it would be easier) (the learning process for this would probably involve a lot of pain unless you had help). To steer the spool, I built up some yaw momentum and then imparted this yaw while raising the inside wheel of the spool.

The most unusual "thing" I rode was the supports used to move trampolines. It's an upside down T with two castered wheels. You rode it like a scooter, but the control issue was that both wheels where castered and free spinning. Direction and lean control were accomplished by leaning the vertical bar on the support side to side. There was no yaw control so if one wheel had more drag the then you had to do a mild weave pattern to make sure your left and right inputs balanced each other out. The wheels had great bearings, and a good initial speed was enough to traverse a very large gym.

I messed around with gymnatics, and was able to do handstands. The easiest is parallel bars, because you can apply a torque force forwards and backwards. You can even lean to the side if the p bars are secured (I only witnessed one guy try this on an unsecured set). The ground is a bit harder as you can only apply significan torque force in one direction. The rings are hardest because there no torque force involved. Instead you have to chase your center of mass by translating the rings with your shoulder muscles forwards and backwards, with a bit of overshoot to correct leans. Handstands on a moving skateboard were similar to p bars, control inputs were the same as when you rode right side up, so virtually no learning curve other than using a cartwheel with 1/2 twise like motion to end the ride on the far side of the skateboard.

The main thing I got out of these experiences was a sense of control inputs and how easy or difficult they would be.

Since I bored the group with this off topic post maybe this short clip of a guy doing a quadruple back flip off the flying rings (no longer a gymnastic event), will help compensate:

http://jeffareid.net/real/quad.wmv

 

[lagoausente]

Hello BigShox, I hope you are the first over the world to implement a real countersteering response, I was searching for that..., the pc sim is...mmmm Rf****r? I hope so..

Look here:
http://perso.wanadoo.es/jcgtortosa/pilotaje/marcoGeneral2.htm
and the pdf that is on that page: http://perso.wanadoo.es/jcgtortosa/pilotaje/manual%20de%20pilotaje.pdf

That web is very good, is in Spanish so if you only speak english maybe google translator can help with copy/paste from pdf.

Look on pagge 40 on the pdf, that explains each of the phases of the counstersteering and lean action.

The body lean is only a complement, mainly for below the center of gravity of the machine-man combo. It´s true that you can lean the byke a bit with the body movements, but only in a low percentage.
The own front wheel countersteering action also afects, but only 10% of total.
What really makes the bike lean and wake up is centrifugal force, this depends on the speed and the curve radius.
What really the pilot does is to vary the curve radius and/or speed, that that´s countersteering.
Example:
Your are in the middle of a curve (turning right), if you turn the handlebar to left, you´ll increase the radius, the bike will lean right.
If you turn the handlebar to right, you´ll decrease the radius of your drawn, that will make the bike get up.
The combo man-bike mass is all the time in balance with the centrifugal force, and this one depends on
1) speed
2) radius

Note that a very small variation in the drawn makes the bike leans or get up, is very sensible. So that´s why the handlebar only need some presure to make big effects on the lean.
I tried some games with a G25 wheel just inverting the wheel polarity, that means, turn left the wheel to lean the bike to right. Setting the G25 to only 40º and max sensibility is posible to game with countersteering... but there is a problem, in real life, when you are on straight and you have to turn right, just need to turn the the handlebar to left "for a moment", just that moment needed for the leaning, but once the inclination is right for that curve the handlebar goes to center, little center-right.
In my trying with the G25 I must keep pushed the G25 to left during the whole curve, what´s not real.

 

[lagoausente]

Thinking more about where the games fail:
Let´s suppose I invert the handlebar, so turning it left, lean the bike right (believe me, I ride bikes). The bug on existing sims is that the bike gets up alone one you put the handlebar to center and that´s not true
Once the bike is inclined will remain inclined because from the moment it leans it starts to draw a curve, with more or less radius depending on the lean.
So control must be:
handlebar left: lean right
handlebar right: lean left
handlebar center: maintain actual lean.

Once you can get this, maybe you can add the man-weight action, but as told, regaring the leaning is only a little % that only affects "down" and never "up". A bike weights 200kg and the reason why the man can change his position so quicly is because the man also is afected for the centrifuge force just like the bike, so in reality, the centrifugal force helps the man, not the man manage the bike with his force (this is a myth).

 

[rcgldr]

The bug on existing sims is that the bike gets up alone one you put the handlebar to center.

If the handlebar represents steering postion, then centered handlebar translates into straight ahead steering, and if the bike is leaning it will fall over on the inside. If the handlebar represents countersteering force, then it depends on the motorcycle, but most motorcycles at reasonable speeds (below 90 mph) tend to straighten up if you relax the force on the handlebars. This is because the steering geometry (trail) is set to cause the front tire to steer inwards a bit more than needed to maintain a lean, a self correction feedback, that results in the bike returning to vertical if no opposing torque to this self-correction is applied to the handlebars.

Can someone blow the dust off this old thread?

 

[lagoausente]

If the handlebar represents steering postion, then centered handlebar translates into straight ahead steering, and if the bike is leaning it will fall over on the inside. If the handlebar represents countersteering force, then it depends on the motorcycle, but most motorcycles at reasonable speeds (below 90 mph) tend to straighten up if you relax the force on the handlebars. This is because the steering geometry (trail) is set to cause the front tire to steer inwards a bit more than needed to maintain a lean, a self correction feedback, that results in the bike returning to vertical if no opposing torque to this self-correction is applied to the handlebars.

Can someone blow the dust off this old thread?

Not agree, my experience on my bike match perfectly with description on the pdf I have linked, the four phase description is perfect. The architecture of the bike is done for a stable equilibration on any lean, the handle and the tyre make the steering turn to the same side of the lean to breakeven.
From page 42:

Después esta velocidad de inclinación se va agotando ‐debido a los mecanismos expuestos anteriormente‐ de forma que cuando llega al ápice, la moto ha dejado de aumentar la tumbada y si aún tiene tendencia a inclinar algo más, el poco de gas que se implementa en la FASE III del giro, la termina de contener, la moto ha alcanzado su nuevo equilibrio. En este momento y si el piloto quisiera, podría soltarse de manos y la moto seguir estable en su trazada (Fig. 5.23): la rueda trasera ha tomado su protagonismo y la fuerza de la gravedad ‐que empuja la moto hacia el asfalto‐ está equilibrada con la fuerza centrífuga ‐que tiende a levantar la moto de la tumbada‐.

Google translate:

Then there is tilt speed is depleted, due to the mechanisms outlined above, so that when it reaches the apex, the bike has steadily increased the lying and if it has a tendency to lean a little more, little gas is deployed PHASE III of the rotation, the ends to hold the bike has reached its new equilibrium. At this time and if the pilot wanted to, could come off of hands and the bike remained stable in their drawn (Fig. 5.23): the rear wheel has taken his role and the force of gravity, which pushes the bike into the asphalt, is balanced with the centrifugal force which tends to lift the bike from the grave..

fig 5.23 atached.

 

[rcgldr]

My experience ... fig 5.23 atached.

Racing motorcyles and some racer replica street motorcycles have less trail than more conventional street motorcycles. In addition, some of them have steering dampers. The combination of less trail and/or higher speeds, results in neutral stability in turns, where the bike tends to hold a lean angle if you relax on the handlebars. On common motorcycles, neutral lean stability occurs at around 100 mph (160 kph) (or faster). At 70 mph or less, common motorcycles tend to straighten up if the rider relaxes on the handlebars.

 

[lagoausente]

Racing motorcyles and some racer replica street motorcycles have less trail than more conventional street motorcycles. In addition, some of them have steering dampers. The combination of less trail and/or higher speeds, results in neutral stability in turns, where the bike tends to hold a lean angle if you relax on the handlebars. On common motorcycles, neutral lean stability occurs at around 100 mph (160 kph) (or faster). At 70 mph or less, common motorcycles tend to straighten up if the rider relaxes on the handlebars.

I´m not sure what are steering dampers on motorcycles. Anyway game simulations often try to sim racing motorcycles, so that should be the desirable behaviour. Relax no lean change, push on handlebar lean countersteering, increase speed decrease lean, decrease speed increase lean. Combo of handlebar/speed should manage the lean. Now the question is; is there any sim over the world that uses the handlebar/lean like the real racing motorcycles?
That´s what I was looking for when I found this post.

 

[rcgldr]

increase speed decrease lean, decrease speed increase lean.

Usually speed changes don't affect lean angle, even if a wheelie is done while leaned over, the bike retains it's lean angle (if it doesn't slide, and rider doesn't compensate prior to the wheelie). Since braking puts a force on the side of the front tire, some motorcycles get an inwards steering torque during braking, and this can reduce lean angle if the rider doesn't compensate with opposing torque inputs, while other bikes (or perhaps the tire profile) seem to not experience this inwards torque effect due to braking.

 

[lagoausente]

Usually speed changes don't affect lean angle, even if a wheelie is done while leaned over, the bike retains it's lean angle (if it doesn't slide, and rider doesn't compensate prior to the wheelie). Since braking puts a force on the side of the front tire, some motorcycles get an inwards steering torque during braking, and this can reduce lean angle if the rider doesn't compensate with opposing torque inputs, while other bikes (or perhaps the tire profile) seem to not experience this inwards torque effect due to braking.

I think you are wrong;
Lean depends mainly on centrifugal force. Gyroscopic efect from wheels helps the bike to maintain stability but the lean changes come from centrifugal force.
The tyres are the supporting point, the drawn they describe are determined by the handlebar. The upper zone of the bike and the rider suffer the centrifugal force, and mainly this one determines the leaning.
At the same time centrigual force is depends on one parameter, the angular speed (or radian frequency as you like).
Note that at a constant radian frequency, the more radius that a circle has, more linear speed is needed, and the less radius less linear speed.
Example: Sweden and Spain have the same angular speed, they complete 360º on 24 hours, but Spain has a higher linear speed then Sweden. As an extreme, a point on the Earth at 1km from the north pole, also turns 360º in 24 hours but near zero linear speed.

That explains that on a easy curve a rider can get the max lean at a much higher speed that on sharp bends, as more sharp, less speed en vice versa.
what I want to say with all of this?
We have two parameters that determines centrifugal force and so the lean:
1) linear speed
2)Radius
At a constant radius if you increase linear speed will increase centrifugal force and so the bike will tend to rise and vice versa.
At constant linear speed, decreasing the radius will increase centrifugal force, increasing the radius will decrease centrifugal force and so the lean. This last one is the "countersteering" effect, turning to left lean to right.
There are a lot other parameters, like the brakes, the countersteering effect of the own front wheel (only 10% of total) , the rider weight changes etc, but that should must be considered as extras, not the mainly behaiviour of the bike.

 

[rcgldr]

Usually speed changes don't affect lean angle ... (other than braking on some motorcycles which causes the front tire to turn inwards).

I think you are wrong; Lean depends mainly on centrifugal force. ... At a constant radius.

Lean angle changes require a torque along the roll axis, which is due to a combination of gravitational and reaction force to centripetal acceleration at the center of mass and the forces (vertical and horizontal (centripetal)) at the contact patch of the tires.

On a motorcycle that tends to hold a lean angle when the rider relaxes on the handlebars, speed changes result in radius changes but the lean angle is essentially held since that is the tendency of such motorcycles.

 

[lagoausente]

Lean angle changes require a torque along the roll axis, which is due to a combination of gravitational and reaction force to centripetal acceleration at the center of mass and the forces (vertical and horizontal (centripetal)) at the contact patch of the tires.

Agree

On a motorcycle that tends to hold a lean angle when the rider relaxes on the handlebars, speed changes result in radius changes but the lean angle is essentially held since that is the tendency of such motorcycles.

Not agree, linear speed changes affect the angular speed but not the circle that the bike is drwaing. (not before lean change)
On bikes you describe,when they are keeping the lean, the handlebar and front wheel gets an angle in respect to the rear by itself. So even the rider relax the torque the bike has there own angle between both wheels. On a curve to right, the handlebar will be a little to the right. The smaller circle (sharper curves) more to right will be the handlebar and vice versa.
On a bigger circle to have the same lean the bike must have more linear speed, on that bigger circle the handlebar will be a little more to center (by itself bike behaviour).
Now suppose we have the bike running at 100kmh at a circle, if you increase the linear speed, centripetal force will increase. If your theory is ok, for the bike keeping the lean must increase the circle radius, and so the handlebar (on a curve to right) will have to go a little to center. But the front wheel gyroscopic effect will avoid it.
The own gyroscopic wheel effect (10% of total of the byke coutersteering effect) makes that if you torque the handlebar to left, then the wheel itself will tend to lean to right (see the pdf linked). So on the situation described, is supposed that on the situation I described before, when linear speed increase, to keep the lean the front wheel must go a little to center (left) so gyroscopic wheel effect will make the wheel lean to right. Since it´s only 10% of total lean force of the bike, the wheel will find some intertia to change it´s actual position. Result must be that the increasing the speed the bike tends to rise up (if no torque is applied by the rider because of the own wheels intertia that offers stability).
Racing experience tells that, look at the pdf:
http://perso.wanadoo.es/jcgtortosa/pilotaje/manual%20de%20pilotaje.pdf
On page 42:

En la FASE IV del giro, momento de aceleración progresiva, con la distancia entre ejes totalmente recuperada y la rueda trasera empujando desde atrás, la tendencia de la moto es a levantarse de la tumbada de manera progresiva. Aún así, si el piloto necesita poner la moto vertical con más rapidez, deberá aplicar contramanillar en el sentido inverso al que lo hizo para entrar a la curva y con la progresividad que requiera la situación.

google translate:

In the Phase IV of the turn, when progressive acceleration, with the wheelbase now fully recovered and the rear wheel by pushing from behind, the tendency of the bike is to get out of lying in a progressive manner. Still, if the pilot needs to get the bike vertical faster, contramanillar be applied in reverse than it did to enter the curve and the escalation that the situation requires.

The pdf is based on years of experience on Racing competition and from page 40 it describes each detail of the bike behaviour regaring countersteering and all the phases before and after the lean. Google translator que give a help.

PD: google translator is not very good: contramanillar means countersteering.

 

[rcgldr]

Linear speed changes affect the angular speed but not the circle that the bike is drawing.

Depends on the motorcycle, some tend to hold lean angle even under acceleration, even if the bike does a wheelie, depending on rider position.

Now suppose we have the bike running at 100kmh at a circle, if you increase the linear speed, centripetal force will increase.

Only if there isn't some inherent tendency for the steering to straighten out the small amount required to hold a lean angle, in response to acceleration, which is a potential response depending on the motorcycle. When braking the force on the side of the contact patch can generate an inwards torque along the steering axis, causing the bike to straighten up; this effect varies between motorcycles and tire profiles.

For a racing motorcycle, the rider is hanging off, and the offset center of mass results in a tendency to lean further inwards, so the rider is never truly relaxing on the handlebars, just reducing the amount of steering input to a very small amount of torque on the handlebars when holding a lean angle.

The own gyroscopic wheel effect (10% of total of the byke coutersteering effect) makes that if you torque the handlebar to left, then the wheel itself will tend to lean to right.

... and the rear tire, will generate a yaw response to the right in response to a lean to the right, creating a small torque to the right at the contact patch of the front tire, somewhat opposing the left torque applied by the rider, but this is a small effect.

That 10% effect depends on the speed. As mentioned, once above some relatively high speed, the gyroscopic effect would tend to cause under-correction, called capsize speed in mathematical models. However, tires have a round profile, and the contact patch being on the side of the tires when leaning counters the capsize factor, and the end result for most bicycles and motorcyles at high speeds is a tendency to remain stable or at least change lean angle at a very slow rate. Camber thrust and torque forces at the contact patches further complicate matters.

Link to wiki article that includes info about capsize speed, but the not the opposing factor from contact patch with a round tire profile:

http://en.wikipedia.org/wiki/Talk:Bicycle_and_motorcycle_dynamics#Eigenvalues