As someone that works with big hydraulic tractors the steering modes have (nick)names and are actually used a lot for different applications. One tractor I drive can do all 3. Front wheel only is called Car or Front wheel only steering Mirrored is called articulated steering and is used for big tractors to turn sharply Copied steering is called crab steering and when implemented properly allows the machine to keep facing what it's working on and but move left, right, forwards and backwards to adjust location without swinging around. Great to see these modes explored elsewhere
Also copied steering is used by tractors and heavy machinery to reduce soil compaction by allowing each wheel to have a driving lane and no lane is passed twice.
veritasium is wrong, steering is not needed to keep upright. a unicycle has no steering wheel yet is very easy to keep upright by literally just balancing. he did a poor demonstration because he used a case to prove himself correct where there was bias TOWARDS his hypothesis: of COURSE no one is able to balance on a bike with locked handlebars, because everyone is already USED TO using the handlebars! how are they supposed to adjust to riding this way in the first 5 minutes that he recorded them? there is a reason why no one can ride a unicycle or EUC in the first 5 minutes either.
@@Blox117 >> a unicycle has no steering wheel yet is very easy to keep upright by literally just balancing. You don't need a steering wheel to steer. I'm the guy that rode the unicycle in the Veritasium video, and he definitely did not get it wrong. >> there is a reason why no one can ride a unicycle or EUC in the first 5 minutes either. I'll tell you what... I'll give you a week to practice on my bike with steering locked. I've got $10K that says you won't be able to ride it at the end of that week.
You should add a gyro and have the back wheel steer towards the tipping side so it helps you keep upright. Also, perhaps a mode which doesn't turn when you're turning slightly, but ramps up and makes sharp turns even sharper.
Instead of a switch to change the steering mode you could install a potentiometer. The middle position is for back wheel steering off, and turning it to the right or left you can adjust the strength of the steering in both modes between 0 and 100% (better not more).
I was going to suggest along the same lines. Though I would set it so on extreme settings the rear wheel turns _more_ than the front wheel. A second potentiometer for fine-tuning the alignment might also be helpful.
Reminds me of the Swing Bike from when I was a kid in the 70s. I remember really wanting one because you could ride with one wheel up on the curb and the other on the street. Now that I’m older they seem like death traps! Lol. There is a video on here demonstrating the original pedal powered swing bike. Look it up.
The velocity should change for each wheel based on the length of path it takes, as curvy is longer than straight. The front bearing of a bike wheel allows for velocity change and I assume it contributes to stability as well.
Hey James! This is somewhat unrelated to the video, but I’ve noticed you use PID in all of your balancing robots, have you considered showing off any other control methods? Such as sliding mode or lyapunov controllers? Would be awesome to show off more methods of robot control Thanks for the awesome videos!
You should look up a swing bike, it’s kind of similar to the second mode but the back wheel doesn’t have forward steering geometry. And the pivot is under the seat so it’s controlled by the rider’s hips and body weight. It’s a sort of novelty/ cruising bike usually you see people ride them in flat areas near the beach like
I think it might also be interesting to have a bit more delay before the back wheel starts to turn (I know there is a tiny lag already) - or just have the back wheel steer more slowly than the front one is turned - this might allow you to lean into the turn a bit, but still execute a small turning circle. Or it might be a disaster.
Currently the thing has lag, as in: the back wheel moves a few tenths of a second after receiving a command from the front wheel. This seems extremely bad because not only does it not steer when your brain tells it to, it continues to steer when it's nolonger wanted. I do agree on making the back wheel steer less or to only start steering after the front wheel has turned a to a particular angle, but that's mainle just for controllability.
What a great video. I really like how each step is covered and how each sponsor was recommended, especially as you've clearly used their parts/machines. It's the best type of recommendation.
Found this channel like 10 mins ago and already love it! Love the similarities of him being a genius in his field with “kid like questions”. Good stuff
This reminds me of Honda's four wheel steering (4WS) which they used on at least 1 production car. While they took an all mechanical approach, I believe you would be able to adapt it quite easily with your knowledge.
Some elastic stirrups could support your legs while traveling, but also allow someone to catch themselves if it tips over. This contraption is one of my favorite. Great video!
I'm pretty sur the laguna 4 wheel drive use both modes, opposite direction for tight turns at slow speed, and same direction for low turn at higher speed, however with same turn direction, they don't rotate by the samùe angle, the rear rotates way less. It is also used by porsche too now
The 4-wheel steering feature was designed for better manoeuvrability at low speed by lowering turning radius and for minimising fishtailing when changing lanes at high-speed. Would be interesting to test those two modes with this 2-wheel steering bike
There's an Irishman who made a bicycle that has both wheels steering. I had a go in Hyde Park, London. It worked very well. Weird, but hella fun. It wasn't electric at all, this was back in maybe 2004. The bicycle had conventional pedals, and a front and rear handlebar that you controlled. One end steered by one arm, the other by the other arm.
AFAIK the Porsche Taycan offers two axes steering as an option and it uses both modes, parallel and opposite direction steering, depending on the current speed. At higher speeds (above 50km/h) it uses parallel steering, which is supposed to improve comfort when overtaking other cars (which is the main purpose of a Porsche). At lower sppeds it uses the rear wheels steer in the opposite direction, which helps for parking.
This build reminds me a lot of that pink panther episode where he builds a bunch of bicycles out of weird shaped parts and struggles all over town on them. Excellent work as always, thanks for the vid. Very entertaining and informative.
I would have expected the parallel steering mode to be about as hard to ride as the normal bike with the steering wheel fixed. I think the reason why this works at all is because the steering of the rear wheel is lagging that of the front wheel due to not being hard-linked by means of a chain but using a servo motor instead.
In parallel steering mode you can't change directions, but you can easily move laterally. That's how you balance an inverted pendulum, and that's nominally how you balance a regular bike.
You could make a "drift" or "ice" mode by making the back wheel steer the opposite direction from the front wheels with a slight delay. Maybe change the amount of steering based on the speed you're driving at as well.
I rode a bike back when I was a teenager in the early 2000s called a swing bike. It had pivot points at the handlebars and at the seat. It allowed the center frame to move and that changed the riding dynamic. It essentially gave the bike all wheel steering. And gave interesting riding possibilities like 1 wheel on the curb and the other wheel in the street
You should try combining your steering modes into a single adaptive mode that scales the back wheel position from a negative multiple of the front wheel to a positive one as you increase in speed. You'll want the entire range somewhere between -1 and 1 so the back wheel always moves less than the front wheel. This will give a very smooth and stable feel where the same input makes you turn more sharply the slower you are going. Some fancy cars have used this idea to make it more comfortable to drive both city and highway.
might be quite good to have those modes as momentary contacts. so you can kind of pop a trick move into a corner like a drift or super turn but as soon as you loose it, its back to regular steering again.
@@Stoneman06660 he needs to get rid of that cheap tig get a proper 1 that will be a good start those scratch start lift tigs your just along for the ride they suck
Very cool experiment with easy to see differences! Reminds me of a certain car made decades ago that came with 4wheel steering, they would turn in the same direction for "lane-change" stability, but some people would modify them to turn in opposite directions for extra quick turning.
I made a 2 wheel steering bike in middle school and it was crude but fun. The seat and rear triangle frame with pedal cassette and rear wheel were welded into the front forks of a second parts bike. The seat was welded in place of the handle bars and this assembly was then welded to the fork, handle bars and front wheel from first parts bike. The bike was a hit everywhere I went. One wheel on the sidewalk and one in the road, riding side saddle was also entertaining and doing super small donuts finished off the trick list. I remember seeing a production looking double steering bike back in late 90’s. It had a spring to help the rear return to center and also had a lockout pin to make it conventional steering.
The steering axis appears to intersect with the wheel hub, so these wheels have no trail/rake. The trail is what makes that automatic steering effect described in the beginning of the video, not the "head tube angle".
James, the good old B-52 aircraft has a steering system that allows the airframe to point in to the wind while the wheels track straight down the runway (crosswind crab). You could do something similar with the bike. Pretty cool, thanks!
Chevy pickups around 2002 had Quadrasteer, a four wheel steering setup. The way they did it was to restrict the crab mode to below a certain speed (about 15 mph) and above that speed they used mirroring mode. They also restricted the rear angle to about 15 degrees in both modes, and the angle was based upon front steering input (a percentage of it). Using crab mode above 15 mph resulted in dangerous unstable conditions (roll overs), and using mirroring below 15 mph made it difficult to park. Using crab mode at below 15 mph made it easier to park and/or maneuver in tight spaces, and using the mirroring mode above 15 mph made lane changes easier. You may find that this method works well on your bike. You probably want your speed limit to be lower than 15mph, but the limit on the rear steering angle should stay around 15 degrees either side from center. I would start testing your speed limit at 5 mph and work from there.
you need something like a peg so you can balance by shuffling your body around when using the two rear steering modes putting a ratio on the rear wheel steering would probably help, i'd want to see a version though where it tries to balance the frame depending on the speed, roll, and steering input angle see how fast you can go and still turn very well. also, since the rear wheel also has positive caster angle, i wonder how a negative caster angle would behave.
That is very impressive. Have you seen the 1988 Honda Prelude Si 4WS? For large steering angles the front and rear wheels steered opposite, and for very small angles they steered the same direction front and rear. Having your front and rear steering different angles will give a greater area of control..
If memory serves me, General Motors had tried active rear wheel assistance steering. The rear wheels turned based on speed. At slower speeds the wheels turn opposite, think tighter turns, smaller turn radius. When the vehicle was moving faster the wheels turned the same direction, think lane change at highway speed.
When the wheels turn in different directions, it accelerates the turn, but the rear wheel makes it difficult to balance. But since the rider knows and expects the turn to occur, it is easy to control. When the wheels turn in the same direction, it makes balancing extremely easy. The bike rides under the center of mass at twice the speed. Probably such an option would be a very convenient training bike for those who do not know how to ride a bike. One or two attempts to keep the balance and spinning the handlebar in a random direction will allow you to almost immediately understand the principle of balancing. (I learned to ride a bike for several years, I know what I'm talking about. Just one tip could teach me in a minute.)
Thanks for doing all of the experiments with steering directions, which is what I always wondering about and wanted to do someday. Keep up the great work.
When you turn on a normal bike the two wheels don't follow the exact same path. The rear wheel would normally be a smaller circle. If that wheel is trying to go the same speed, that force would need to do something. I wonder if that is why the zip ties broke and why the tracking seems off. You should try to look at the tracking of the wheels on the ground. A hard packed dirt path would probably work to leave a visible path.
This gives me two ideas. Steering of cars, vans and lorries is mechanical; the driver turning the steering wheel determines what the wheels do. But when a vehicle goes round a curve, the wheels REALLY ought to be directed along a tangent to that curve, whereas in most vehicles, the wheels that steer (unfortunately, also called steering wheels; don’t confuse them with the driver’s steering wheel) are all parallel to each other, and the non-steering wheels are all parallel to the vehicle. What if you could steer all the wheels electronically, based on how far round the driver turned the steering wheel: the further round, the smaller the radius of the curve you want to turn. If the front two wheels are angled at 2º to the right and the back are angled at 1º to the right, that puts you on a curve of about radius 115w, where w is the wheelbase. If the front two wheels are angled at 2º to the right and the back are not angled at all, it’s about 57w, and if the back are at 2º to the left, it’s about 29w. If the radius is zero, the wheels form a circle and the vehicle spins, like an engine on a turntable. If the vehicle wants to park in a small space, it can set its wheels at 90º to its body and go sideways. The other idea is to do with back wheels “tracking” the front wheels. Consider a lorry tractor with two trailers. The driver turns a corner; the first trailer follows it round fairly accurately, and the second will slightly tend to cut the corner. Now consider having ten trailers: shorter, less weight. The third would cut the corner worse, and the tenth would far worse. But if their wheels were steered electronically, and told to go where the tractor’s went, they wouldn’t. And if you could offload some of the tractive force of the tractor to the trailers, no axle would be very heavy and you could carry more, driven by one driver. Turning round, I haven’t yet thought of, and dealing with them loading and unloading. Instead of one lorry and a 40ft container, one lorry could pull a long line of 10ft containers . . . AH! I’ve just been googling, and found this: www.researchgate.net/figure/a-A-kinematic-model-of-the-truck-carrying-the-wing-the-trailer-is-hitched-to-the-truck_fig7_3344095
Hi quick question at 3:18, what type of bearing in the one be pushed on the table, and moment later the one on the right, out of the two on the screen, I have portable bag-less vacuum cleaner, and it has something the same that part if getting dust collected, what ever it sucks up, as part, and it needing regular cleanout (I double, as part it was built to last, so now what it called would be handy to now?
The bearing on the left, the one he presses his palm on, is a thrust bearing. It's designed to take primarily axial loads. The bearing on the right is a pretty typical thin-race bearing with steel dust shields. Traditional bearings are designed primarily for radial loads, but can take some axial load. If you want a traditional bearing that handles more axial load, opt for a deep-groove bearing.
This is cool! The only thing I would change is in programming. I would have the rear wheel rotate at a percentage of the front like half or 2/3 of the front input. That would allow you to still have directional control.
There is one car (German made, i think. Because of course.) that uses the "crab steering" at high speeds to make lane changes on the highway smoother, and the "counter turning" steering (where front and back wheels turn in opposite directions) at the very low speeds for better maneuvering in parking lots or other tight spaces. It uses normal "front wheels only" steering for everything between those two extremes.
You could motorize the front steering like the back, and then have both steer by half as much when you are in the mode where it felt weird because it was steering too fast.
My control systems professor did a whole lecture on bicycle dynamics. An interesting point that came up was that it is possible to make an unridable (at least without some control system) bike that is steered by the back wheel. This might be fun to check out as well
I made an electric trike with front drive and rear steering, one driven front wheel, two undriven rear steering wheels. it had a keen desire to eject the rider without constant correction.
One thing you could try is something I dimly recall from experimental cars with "parking assistance": when you turn the handlebar, the back wheel initially turns in the same direction as the front wheel, but as you turn the handlebar more the back wheel changes to turn in the opposite direction until it catches up. I think that translates to your mode 2 for small turns but your mode 1 for bigger turns, with some tuning required for the transition no doubt ;-) IIRC the idea was that this would help with parallel parking, and changing lanes: I have no idea whether it worked but it stuck in my memory!
This actually seems like it could be a useful product to actually exist. An e bike with additional modes like this would be great for better turning and trandporting items!
I love it! I want to make a prediction before watching... I think it will be relatively easy to balance on the bike when both wheels steer in the same direction. I suspect you'll be able to learn to ride it when they steer in opposite directions, but I expect it to be more difficult. I'll grade myself in a few minutes after I watch.
I think I'm going to give myself a C for those predictions. I figured with the wheels turning in opposite directions you'd have the following issue... The front wheel would operate exactly as an inverted pendulum. Steering into the fall will tend to "catch" you. But the rear wheel will be trying to steer out from under you. So the immediate response will do very little to help you balance. But of course the slightly slower response is that the bike will change direction in the desired manner, and that will allow you to balance once you get used to it. With both wheels turning in the same direction, the bike itself would never change directions. But you can make it translate immediately to the left or right, thereby allowing you to balance as an inverted pendulum. I still believe that's the case. But I think you had two small but significant problems with the build. As you noticed, the rear wheel was not steering the same amount as the front. Also, It looked to me like there was a small amount of lag between the front wheel and real wheel steering. In any case, very nicely done. By the way, I'm the guy the modified the bike for the Veritasium video.
I think it was Honda had an adaptive 4wheel steering option that would vary the steering mode depending on the speed of the vehicle and the radius of the turn. So as you cranked the wheel harder at slow speed, it would go from front wheel steering to all wheel steering. And if you were at highway speeds it would start out in parallel steering and switch to front steering if you turned farther.
Do you have any videos that show how to wire and run the hover board motor? Bts7960 only has two wires that go to the motor, but the motor has 5+3 wires.
Another channel modified a bike so the handlebars were disigned to steer in the opposite direction, so when steering the handlebars to the left the front wheel turned to the right. It was like learning to ride a unicycle it was so difficult to ride.
Looks like mode 2 (crab steering, two wheels turning the same direction) is helped by the delay between the front and rear wheel. It appears you're correcting the lean quickly enough so that the rear wheel doesn't have time to turn, effectively making it a front-wheel-steering bike for just a fraction of a second but enough that you can balance. I imagine this is why it doesn't work well at slower speeds, since at slower speeds steering inputs need to be slower, giving the rear wheel a chance to catch up to the front.
Nice experiment! Looks like the rear steering is a lot slower than the front steering and that's messing up your balance a lot. A re-configurable mechanical linkage between the front and rear steering might give better results?
James your works are very 'professional' looking and if only that was possible for me, but for anyone who might be wanting to try out different systems etc, I built a rough wooden try out of a bike from old scrap pieces and used a mechanical link to connect up the wheels so they would do the same experiment as what you've done. And I got the same results. The cost of my simple version was about £5 or so in materials but several days labour to get it all together. But James, very professional yeah, but what are you gonna do with it all now. My rough wooden try out has gone back to nature, can you say as much?
So how would the front wheel locked out apply to your previous build of the self balancing bicycle. With self balance mode turned on and the front wheel locked out, can the bicycle still balance itself?
I wonder if the steering could be used as the on/off switch once it hits a certain point? Would it then act like a drift in a corner and then go back to normal once you pull out of the turn
I think you were only able to ride the bike in that third mode because the steering motor on the rear wheel acts slower than your manual input on the front. Steering the front wheel into the tilt helps you stay upright. The rear wheel steers against the tilt and tries to throw you off balance. But since the rear wheel steers so slow you are able to get some microcorrections in at the front without the rear moving much. This is saving that little bit of balance you have. If the wheels were truly steering identical, either with both of them being mechanically linked or both driven electricly at the same speed, that third mode would be unridable.
For the normal two wheel steering mode, you need the back wheel to not turn as much as the front wheel. Ramp up the amount of back wheel steer as you turn the front wheel further. That way your small correction moves aren't amplified but you still get tight turns.
Lovely work Sir Bruton. I suggest you make this idea using standard bicycle steel cables and some pullies for steering both wheels without batteries and motors, you might avoid most of the problems you had in this version.
What about something like this that transitions from being a hoverboard to being a two-wheeled self-balancing longboard? So it can do the self balancing thing with the wheels perpendicular to the board but then the wheels themselves can rotate like casters to be parallel to the board and then you're just standing on a very short bike, which would then try to steer into turns to keep you upright.
The conventional rear wheel steering would be far more usable if the rear wheel did something like half the movement of the front wheel - very few rear steering vehicles match the steering angle as it causes the back to swing out too much, as you found.
Go full steer by wire and have the yaw angle determine the turning radius and use a secondary control like twist grip or handbrake to vary the steering contribution between front and rear wheel
There are some interesting parameters you could tune in this setup. I.e. it might be interesting to see what happens if you change the ratio between the front and rear steering angle
Is there a chance that you are using the misalignment of the wheels to still turn and thereby balance yourself? I know from my gym that the bidet always try to find the easiest path to complete a task even if you don’t actively think about it. So maybe make sure first the wheels are perfectly aligned then do the same test to confirm.
I would like to recommend two things to improve this, and with (1. Foot Rests) you can swing the bike like a swing bike and (2. check out some videos on TH-cam for swing bike). Basically you use your feet on the pedals to swing out the frame, and it will make both alternate modes more useable.
I'm sure others suggested it, but you could try to reduce the steering on the rear wheel a little bit. That way it should adjust the sideways thing to come back into a straight line, but keep the basic idea of steering both wheels the same way.
why did you do it via software? you already have two pullies on the steering wheels, just put a long belt over both, and flip the belt once on the back to change the steering mode
what if you made the rear wheel steer independently of the front - like maybe have a lever you can activate with your right hand or your feet so you can control when you're "crabbing" and when you're turning tighter?
Swing bikes were originally a thing in the 70's . my older brother had one and I rode it around for fun in the 80's . it looked like an old stingray but had a knob on the upper tube that you could pull to release the back end and a spring that pulled it towards the center unless you actally pushed the bcak out.
The reason you can ride the bike at the end is the hysteresis or time delay of the rear steering. I would make a simple and crossed chain linkage between the front and back axes and lock and unlock the respective wheels. An H frame with no struts is ridiculous.
I think maybe the only reason you can do it is weight distribution changes to angular. but the cool thing is you can delay the rear wheel to trail the front like this. so you could use mode 1 to stear normal and gradually increase the angle of the rear during a turn making it tighter and tighter. not sure of the application but it would be intriguing to put a decaying delay on the rear angle. kind of interesting concept for a 2 wheel vehicle.
I theorize that the only reason you can balance with the same direction steering is due to the lag between when the rear steers compared to the front. If the wheels were fixed to steer at the exact same time I don't think you would be able to maintain balance.
This is very interesting. Since you have the manufacturing capability, i'd actually be interested in a project where you made a sensible, practical e-bike, maybe suitable for off-road use, and with some handy electronically-controlled refinements that commercial models lack, like for example some sort of active steering or active suspension.
As someone that works with big hydraulic tractors the steering modes have (nick)names and are actually used a lot for different applications. One tractor I drive can do all 3.
Front wheel only is called Car or Front wheel only steering
Mirrored is called articulated steering and is used for big tractors to turn sharply
Copied steering is called crab steering and when implemented properly allows the machine to keep facing what it's working on and but move left, right, forwards and backwards to adjust location without swinging around.
Great to see these modes explored elsewhere
More like snake mode and sideways mode for that goofy bike.
Also copied steering is used by tractors and heavy machinery to reduce soil compaction by allowing each wheel to have a driving lane and no lane is passed twice.
veritasium is wrong, steering is not needed to keep upright. a unicycle has no steering wheel yet is very easy to keep upright by literally just balancing. he did a poor demonstration because he used a case to prove himself correct where there was bias TOWARDS his hypothesis: of COURSE no one is able to balance on a bike with locked handlebars, because everyone is already USED TO using the handlebars! how are they supposed to adjust to riding this way in the first 5 minutes that he recorded them? there is a reason why no one can ride a unicycle or EUC in the first 5 minutes either.
@@Blox117
>> a unicycle has no steering wheel yet is very easy to keep upright by literally just balancing.
You don't need a steering wheel to steer. I'm the guy that rode the unicycle in the Veritasium video, and he definitely did not get it wrong.
>> there is a reason why no one can ride a unicycle or EUC in the first 5 minutes either.
I'll tell you what... I'll give you a week to practice on my bike with steering locked. I've got $10K that says you won't be able to ride it at the end of that week.
@@alexandrsoldiernetizen162 They're actually both snake mode, ever seen a sidewinder?
You should add a gyro and have the back wheel steer towards the tipping side so it helps you keep upright. Also, perhaps a mode which doesn't turn when you're turning slightly, but ramps up and makes sharp turns even sharper.
That would make this so much easier to ride, a rc helicopter/drift car gyro would work perfectly
so cool
these ideas are on point
Instead of a switch to change the steering mode you could install a potentiometer. The middle position is for back wheel steering off, and turning it to the right or left you can adjust the strength of the steering in both modes between 0 and 100% (better not more).
I was going to suggest along the same lines. Though I would set it so on extreme settings the rear wheel turns _more_ than the front wheel. A second potentiometer for fine-tuning the alignment might also be helpful.
With some balance controls you could even use it as a clutch that allows you to drift, which would be a blast to ride I would think!
Reminds me of the Swing Bike from when I was a kid in the 70s. I remember really wanting one because you could ride with one wheel up on the curb and the other on the street. Now that I’m older they seem like death traps! Lol. There is a video on here demonstrating the original pedal powered swing bike. Look it up.
How the heck do you build stuff so quickly??? It’s incredible.
An incredible amount of determination and talent with nothing else to do
He's obviously an android
It’s all done with smoke and mirrors 😁
It’s one of TH-cam’s unsolved mysteries
It's the 1.2mm nozzle that is the secret 😊
The velocity should change for each wheel based on the length of path it takes, as curvy is longer than straight. The front bearing of a bike wheel allows for velocity change and I assume it contributes to stability as well.
Hey James! This is somewhat unrelated to the video, but I’ve noticed you use PID in all of your balancing robots, have you considered showing off any other control methods? Such as sliding mode or lyapunov controllers? Would be awesome to show off more methods of robot control
Thanks for the awesome videos!
Yeah, even a LQR would be good I think.
You should look up a swing bike, it’s kind of similar to the second mode but the back wheel doesn’t have forward steering geometry. And the pivot is under the seat so it’s controlled by the rider’s hips and body weight. It’s a sort of novelty/ cruising bike usually you see people ride them in flat areas near the beach like
I think it might also be interesting to have a bit more delay before the back wheel starts to turn (I know there is a tiny lag already) - or just have the back wheel steer more slowly than the front one is turned - this might allow you to lean into the turn a bit, but still execute a small turning circle. Or it might be a disaster.
Hey didn't expect you here but now that I think about it, it makes perfect sense!
Currently the thing has lag, as in: the back wheel moves a few tenths of a second after receiving a command from the front wheel. This seems extremely bad because not only does it not steer when your brain tells it to, it continues to steer when it's nolonger wanted.
I do agree on making the back wheel steer less or to only start steering after the front wheel has turned a to a particular angle, but that's mainle just for controllability.
What a great video. I really like how each step is covered and how each sponsor was recommended, especially as you've clearly used their parts/machines. It's the best type of recommendation.
Found this channel like 10 mins ago and already love it! Love the similarities of him being a genius in his field with “kid like questions”. Good stuff
This reminds me of Honda's four wheel steering (4WS) which they used on at least 1 production car. While they took an all mechanical approach, I believe you would be able to adapt it quite easily with your knowledge.
There were more 4 wheel steering cars by several Japanese manufacturers
Some elastic stirrups could support your legs while traveling, but also allow someone to catch themselves if it tips over. This contraption is one of my favorite. Great video!
I was not expecting you to be able to balance it when I saw how much lag there was!
I'm pretty sur the laguna 4 wheel drive use both modes, opposite direction for tight turns at slow speed, and same direction for low turn at higher speed, however with same turn direction, they don't rotate by the samùe angle, the rear rotates way less.
It is also used by porsche too now
The 4-wheel steering feature was designed for better manoeuvrability at low speed by lowering turning radius and for minimising fishtailing when changing lanes at high-speed. Would be interesting to test those two modes with this 2-wheel steering bike
@@bornach Agree. The thing to try here is the less rotation on the rear in same direction mode, instead of doing the same amount of rotation.
@@skipspik1571 James Bruton could also try crab walk mode and challenge Derick [Veritaium] or Destin [Smarter Everyday] to ride it.
There's an Irishman who made a bicycle that has both wheels steering. I had a go in Hyde Park, London. It worked very well. Weird, but hella fun. It wasn't electric at all, this was back in maybe 2004. The bicycle had conventional pedals, and a front and rear handlebar that you controlled. One end steered by one arm, the other by the other arm.
AFAIK the Porsche Taycan offers two axes steering as an option and it uses both modes, parallel and opposite direction steering, depending on the current speed. At higher speeds (above 50km/h) it uses parallel steering, which is supposed to improve comfort when overtaking other cars (which is the main purpose of a Porsche). At lower sppeds it uses the rear wheels steer in the opposite direction, which helps for parking.
This build reminds me a lot of that pink panther episode where he builds a bunch of bicycles out of weird shaped parts and struggles all over town on them.
Excellent work as always, thanks for the vid. Very entertaining and informative.
I would have expected the parallel steering mode to be about as hard to ride as the normal bike with the steering wheel fixed. I think the reason why this works at all is because the steering of the rear wheel is lagging that of the front wheel due to not being hard-linked by means of a chain but using a servo motor instead.
In parallel steering mode you can't change directions, but you can easily move laterally. That's how you balance an inverted pendulum, and that's nominally how you balance a regular bike.
You could make a "drift" or "ice" mode by making the back wheel steer the opposite direction from the front wheels with a slight delay. Maybe change the amount of steering based on the speed you're driving at as well.
I rode a bike back when I was a teenager in the early 2000s called a swing bike. It had pivot points at the handlebars and at the seat. It allowed the center frame to move and that changed the riding dynamic. It essentially gave the bike all wheel steering. And gave interesting riding possibilities like 1 wheel on the curb and the other wheel in the street
You should try combining your steering modes into a single adaptive mode that scales the back wheel position from a negative multiple of the front wheel to a positive one as you increase in speed. You'll want the entire range somewhere between -1 and 1 so the back wheel always moves less than the front wheel. This will give a very smooth and stable feel where the same input makes you turn more sharply the slower you are going. Some fancy cars have used this idea to make it more comfortable to drive both city and highway.
might be quite good to have those modes as momentary contacts. so you can kind of pop a trick move into a corner like a drift or super turn but as soon as you loose it, its back to regular steering again.
Your welds look mint, James. Great stuff.
Not mint but do the job
Yeah, commented too soon, @@BIGSMOKE-bl2lq! Some of the later ones weren't grand. Still, leagues better than mine! 😅
@@Stoneman06660 he needs to get rid of that cheap tig get a proper 1 that will be a good start those scratch start lift tigs your just along for the ride they suck
Very cool experiment with easy to see differences! Reminds me of a certain car made decades ago that came with 4wheel steering, they would turn in the same direction for "lane-change" stability, but some people would modify them to turn in opposite directions for extra quick turning.
I made a 2 wheel steering bike in middle school and it was crude but fun. The seat and rear triangle frame with pedal cassette and rear wheel were welded into the front forks of a second parts bike. The seat was welded in place of the handle bars and this assembly was then welded to the fork, handle bars and front wheel from first parts bike. The bike was a hit everywhere I went. One wheel on the sidewalk and one in the road, riding side saddle was also entertaining and doing super small donuts finished off the trick list.
I remember seeing a production looking double steering bike back in late 90’s. It had a spring to help the rear return to center and also had a lockout pin to make it conventional steering.
I can't imagine what it's like to have a crazy idea in your head then make it.
You're like a musician.
The steering axis appears to intersect with the wheel hub, so these wheels have no trail/rake. The trail is what makes that automatic steering effect described in the beginning of the video, not the "head tube angle".
James, the good old B-52 aircraft has a steering system that allows the airframe to point in to the wind while the wheels track straight down the runway (crosswind crab). You could do something similar with the bike. Pretty cool, thanks!
Chevy pickups around 2002 had Quadrasteer, a four wheel steering setup. The way they did it was to restrict the crab mode to below a certain speed (about 15 mph) and above that speed they used mirroring mode. They also restricted the rear angle to about 15 degrees in both modes, and the angle was based upon front steering input (a percentage of it). Using crab mode above 15 mph resulted in dangerous unstable conditions (roll overs), and using mirroring below 15 mph made it difficult to park. Using crab mode at below 15 mph made it easier to park and/or maneuver in tight spaces, and using the mirroring mode above 15 mph made lane changes easier. You may find that this method works well on your bike. You probably want your speed limit to be lower than 15mph, but the limit on the rear steering angle should stay around 15 degrees either side from center. I would start testing your speed limit at 5 mph and work from there.
you need something like a peg so you can balance by shuffling your body around when using the two rear steering modes
putting a ratio on the rear wheel steering would probably help, i'd want to see a version though where it tries to balance the frame depending on the speed, roll, and steering input angle see how fast you can go and still turn very well.
also, since the rear wheel also has positive caster angle, i wonder how a negative caster angle would behave.
That is very impressive. Have you seen the 1988 Honda Prelude Si 4WS? For large steering angles the front and rear wheels steered opposite, and for very small angles they steered the same direction front and rear. Having your front and rear steering different angles will give a greater area of control..
If memory serves me, General Motors had tried active rear wheel assistance steering. The rear wheels turned based on speed. At slower speeds the wheels turn opposite, think tighter turns, smaller turn radius. When the vehicle was moving faster the wheels turned the same direction, think lane change at highway speed.
When the wheels turn in different directions, it accelerates the turn, but the rear wheel makes it difficult to balance. But since the rider knows and expects the turn to occur, it is easy to control. When the wheels turn in the same direction, it makes balancing extremely easy. The bike rides under the center of mass at twice the speed. Probably such an option would be a very convenient training bike for those who do not know how to ride a bike. One or two attempts to keep the balance and spinning the handlebar in a random direction will allow you to almost immediately understand the principle of balancing. (I learned to ride a bike for several years, I know what I'm talking about. Just one tip could teach me in a minute.)
Of course, I'm not talking about the bike in the video. The lack of synchrony and differences in steering angle will make riding a survival test.
Thanks for doing all of the experiments with steering directions, which is what I always wondering about and wanted to do someday. Keep up the great work.
When you turn on a normal bike the two wheels don't follow the exact same path. The rear wheel would normally be a smaller circle. If that wheel is trying to go the same speed, that force would need to do something. I wonder if that is why the zip ties broke and why the tracking seems off.
You should try to look at the tracking of the wheels on the ground. A hard packed dirt path would probably work to leave a visible path.
5:39 did you think to cool the metal off to shrink it just a tad before you fitted the plastic parts onto them?
This gives me two ideas. Steering of cars, vans and lorries is mechanical; the driver turning the steering wheel determines what the wheels do. But when a vehicle goes round a curve, the wheels REALLY ought to be directed along a tangent to that curve, whereas in most vehicles, the wheels that steer (unfortunately, also called steering wheels; don’t confuse them with the driver’s steering wheel) are all parallel to each other, and the non-steering wheels are all parallel to the vehicle. What if you could steer all the wheels electronically, based on how far round the driver turned the steering wheel: the further round, the smaller the radius of the curve you want to turn.
If the front two wheels are angled at 2º to the right and the back are angled at 1º to the right, that puts you on a curve of about radius 115w, where w is the wheelbase.
If the front two wheels are angled at 2º to the right and the back are not angled at all, it’s about 57w, and if the back are at 2º to the left, it’s about 29w.
If the radius is zero, the wheels form a circle and the vehicle spins, like an engine on a turntable. If the vehicle wants to park in a small space, it can set its wheels at 90º to its body and go sideways.
The other idea is to do with back wheels “tracking” the front wheels. Consider a lorry tractor with two trailers. The driver turns a corner; the first trailer follows it round fairly accurately, and the second will slightly tend to cut the corner. Now consider having ten trailers: shorter, less weight. The third would cut the corner worse, and the tenth would far worse. But if their wheels were steered electronically, and told to go where the tractor’s went, they wouldn’t. And if you could offload some of the tractive force of the tractor to the trailers, no axle would be very heavy and you could carry more, driven by one driver. Turning round, I haven’t yet thought of, and dealing with them loading and unloading. Instead of one lorry and a 40ft container, one lorry could pull a long line of 10ft containers . . . AH! I’ve just been googling, and found this: www.researchgate.net/figure/a-A-kinematic-model-of-the-truck-carrying-the-wing-the-trailer-is-hitched-to-the-truck_fig7_3344095
Hi quick question at 3:18, what type of bearing in the one be pushed on the table, and moment later the one on the right, out of the two on the screen, I have portable bag-less vacuum cleaner, and it has something the same that part if getting dust collected, what ever it sucks up, as part, and it needing regular cleanout (I double, as part it was built to last, so now what it called would be handy to now?
The bearing on the left, the one he presses his palm on, is a thrust bearing. It's designed to take primarily axial loads. The bearing on the right is a pretty typical thin-race bearing with steel dust shields. Traditional bearings are designed primarily for radial loads, but can take some axial load. If you want a traditional bearing that handles more axial load, opt for a deep-groove bearing.
This is cool! The only thing I would change is in programming. I would have the rear wheel rotate at a percentage of the front like half or 2/3 of the front input. That would allow you to still have directional control.
There is one car (German made, i think. Because of course.) that uses the "crab steering" at high speeds to make lane changes on the highway smoother, and the "counter turning" steering (where front and back wheels turn in opposite directions) at the very low speeds for better maneuvering in parking lots or other tight spaces.
It uses normal "front wheels only" steering for everything between those two extremes.
14:30 You couldn't turn in that mode just by leaning, while keeping the handlebars straight?
You could motorize the front steering like the back, and then have both steer by half as much when you are in the mode where it felt weird because it was steering too fast.
You got it right. The ones that suggested the handle bars move half as much missed the mark!
You should make it with a reduction on the steering bar so the front fork turns half the turn you're making with the handle bar.
I could do that in code
My control systems professor did a whole lecture on bicycle dynamics. An interesting point that came up was that it is possible to make an unridable (at least without some control system) bike that is steered by the back wheel. This might be fun to check out as well
I made an electric trike with front drive and rear steering, one driven front wheel, two undriven rear steering wheels. it had a keen desire to eject the rider without constant correction.
@@alexandrsoldiernetizen162 th-cam.com/video/x-Blm6gh0Gs/w-d-xo.html
One thing you could try is something I dimly recall from experimental cars with "parking assistance": when you turn the handlebar, the back wheel initially turns in the same direction as the front wheel, but as you turn the handlebar more the back wheel changes to turn in the opposite direction until it catches up. I think that translates to your mode 2 for small turns but your mode 1 for bigger turns, with some tuning required for the transition no doubt ;-)
IIRC the idea was that this would help with parallel parking, and changing lanes: I have no idea whether it worked but it stuck in my memory!
Parallel parking an ugly bike?
@@alexandrsoldiernetizen162 no, parallel parking the car to which this was originally applied
Do you have a mechanical analysis of this two wheel steering bike, it is important to me, thank you
kinematic analysis
I’d love to see a part 2 of this video! This build is intriguing and amazing
That was fun to watch. How about locking the front wheel and steering with only the rear?
Adding a back wheel mode that tries to stabilize the rider upright, and another mode that does the opposite would be fun.
Where do I start looking for an introductory welding course?
could you do just rear wheel stearing?
like a forklift
Hi James I just bought a dual steering bike in the Netherlands. How do I send you a picture?
This actually seems like it could be a useful product to actually exist. An e bike with additional modes like this would be great for better turning and trandporting items!
I love it! I want to make a prediction before watching...
I think it will be relatively easy to balance on the bike when both wheels steer in the same direction. I suspect you'll be able to learn to ride it when they steer in opposite directions, but I expect it to be more difficult.
I'll grade myself in a few minutes after I watch.
I think I'm going to give myself a C for those predictions. I figured with the wheels turning in opposite directions you'd have the following issue...
The front wheel would operate exactly as an inverted pendulum. Steering into the fall will tend to "catch" you. But the rear wheel will be trying to steer out from under you. So the immediate response will do very little to help you balance. But of course the slightly slower response is that the bike will change direction in the desired manner, and that will allow you to balance once you get used to it.
With both wheels turning in the same direction, the bike itself would never change directions. But you can make it translate immediately to the left or right, thereby allowing you to balance as an inverted pendulum. I still believe that's the case. But I think you had two small but significant problems with the build. As you noticed, the rear wheel was not steering the same amount as the front. Also, It looked to me like there was a small amount of lag between the front wheel and real wheel steering.
In any case, very nicely done. By the way, I'm the guy the modified the bike for the Veritasium video.
I think it was Honda had an adaptive 4wheel steering option that would vary the steering mode depending on the speed of the vehicle and the radius of the turn. So as you cranked the wheel harder at slow speed, it would go from front wheel steering to all wheel steering. And if you were at highway speeds it would start out in parallel steering and switch to front steering if you turned farther.
Renault also have something like that if I remember correctly
Do you have any videos that show how to wire and run the hover board motor? Bts7960 only has two wires that go to the motor, but the motor has 5+3 wires.
Another channel modified a bike so the handlebars were disigned to steer in the opposite direction, so when steering the handlebars to the left the front wheel turned to the right. It was like learning to ride a unicycle it was so difficult to ride.
Looks like mode 2 (crab steering, two wheels turning the same direction) is helped by the delay between the front and rear wheel. It appears you're correcting the lean quickly enough so that the rear wheel doesn't have time to turn, effectively making it a front-wheel-steering bike for just a fraction of a second but enough that you can balance. I imagine this is why it doesn't work well at slower speeds, since at slower speeds steering inputs need to be slower, giving the rear wheel a chance to catch up to the front.
Also, subscribed!
Nice experiment! Looks like the rear steering is a lot slower than the front steering and that's messing up your balance a lot. A re-configurable mechanical linkage between the front and rear steering might give better results?
James your works are very 'professional' looking and if only that was possible for me, but for anyone who might be wanting to try out different systems etc, I built a rough wooden try out of a bike from old scrap pieces and used a mechanical link to connect up the wheels so they would do the same experiment as what you've done. And I got the same results. The cost of my simple version was about £5 or so in materials but several days labour to get it all together. But James, very professional yeah, but what are you gonna do with it all now. My rough wooden try out has gone back to nature, can you say as much?
So how would the front wheel locked out apply to your previous build of the self balancing bicycle. With self balance mode turned on and the front wheel locked out, can the bicycle still balance itself?
I like how you used electronics to engineer the steering mechanism rather than just a simple mechanical link.
What will happen if you put a caster wheel at rear, how human steering at front n mechanical of rear caster behave like?
I wonder if the steering could be used as the on/off switch once it hits a certain point? Would it then act like a drift in a corner and then go back to normal once you pull out of the turn
I think you were only able to ride the bike in that third mode because the steering motor on the rear wheel acts slower than your manual input on the front.
Steering the front wheel into the tilt helps you stay upright. The rear wheel steers against the tilt and tries to throw you off balance. But since the rear wheel steers so slow you are able to get some microcorrections in at the front without the rear moving much. This is saving that little bit of balance you have.
If the wheels were truly steering identical, either with both of them being mechanically linked or both driven electricly at the same speed, that third mode would be unridable.
I find the way you start your TIG welds interesting. What are you doing with the torch?
For the normal two wheel steering mode, you need the back wheel to not turn as much as the front wheel. Ramp up the amount of back wheel steer as you turn the front wheel further. That way your small correction moves aren't amplified but you still get tight turns.
Lovely work Sir Bruton. I suggest you make this idea using standard bicycle steel cables and some pullies for steering both wheels without batteries and motors, you might avoid most of the problems you had in this version.
What about something like this that transitions from being a hoverboard to being a two-wheeled self-balancing longboard? So it can do the self balancing thing with the wheels perpendicular to the board but then the wheels themselves can rotate like casters to be parallel to the board and then you're just standing on a very short bike, which would then try to steer into turns to keep you upright.
The conventional rear wheel steering would be far more usable if the rear wheel did something like half the movement of the front wheel - very few rear steering vehicles match the steering angle as it causes the back to swing out too much, as you found.
Go full steer by wire and have the yaw angle determine the turning radius and use a secondary control like twist grip or handbrake to vary the steering contribution between front and rear wheel
i cant say how happy i am to see you not using one of those step "drill bits". im sure theyre great for counter-sinking and precisely nothing else.
James how about a vid on 3d printing tips? Any maintenance you do on your lulzbots, keeping filament dry, design tips?
Your welds are getting much better!
There are some interesting parameters you could tune in this setup. I.e. it might be interesting to see what happens if you change the ratio between the front and rear steering angle
That's quite the surprise, I'd expect the same direction steering to be impossible to ride.
Is there a chance that you are using the misalignment of the wheels to still turn and thereby balance yourself?
I know from my gym that the bidet always try to find the easiest path to complete a task even if you don’t actively think about it.
So maybe make sure first the wheels are perfectly aligned then do the same test to confirm.
I did that as best I could after fixing the broken bits, so it started going straighter after that.
That surprised me. I thought you’d be entirely unable to balance in the third mode. I have been enlightened. Thanks.
I would like to recommend two things to improve this, and with (1. Foot Rests) you can swing the bike like a swing bike and (2. check out some videos on TH-cam for swing bike). Basically you use your feet on the pedals to swing out the frame, and it will make both alternate modes more useable.
I'm sure others suggested it, but you could try to reduce the steering on the rear wheel a little bit. That way it should adjust the sideways thing to come back into a straight line, but keep the basic idea of steering both wheels the same way.
Are you sure you have polarity right on your tig welder? Something seems wrong with your start
why did you do it via software? you already have two pullies on the steering wheels, just put a long belt over both, and flip the belt once on the back to change the steering mode
Hard to do while you're riding.
You should look into recreating the mechanism seen in a coaxial swerve drive - might open up some possibilities for advanced control
what if you made the rear wheel steer independently of the front - like maybe have a lever you can activate with your right hand or your feet so you can control when you're "crabbing" and when you're turning tighter?
My initial thoughts are - the size of the wheel is also responsible for balance - with small wheels you remove that factor.
This is the most over engineered experimental bike I’ve seen in my life and I love it
Swing bikes were originally a thing in the 70's . my older brother had one and I rode it around for fun in the 80's . it looked like an old stingray but had a knob on the upper tube that you could pull to release the back end and a spring that pulled it towards the center unless you actally pushed the bcak out.
The reason you can ride the bike at the end is the hysteresis or time delay of the rear steering. I would make a simple and crossed chain linkage between the front and back axes and lock and unlock the respective wheels. An H frame with no struts is ridiculous.
I think maybe the only reason you can do it is weight distribution changes to angular. but the cool thing is you can delay the rear wheel to trail the front like this. so you could use mode 1 to stear normal and gradually increase the angle of the rear during a turn making it tighter and tighter. not sure of the application but it would be intriguing to put a decaying delay on the rear angle. kind of interesting concept for a 2 wheel vehicle.
One of TH-cam unsolved mysteries. How you have the time to do so many projects back to back
What about using this on a scooter platform?
I theorize that the only reason you can balance with the same direction steering is due to the lag between when the rear steers compared to the front. If the wheels were fixed to steer at the exact same time I don't think you would be able to maintain balance.
Dit you put any offset on the front wheel axle? Would this one self steer when pushed off a hill?
He had offset on both axles.
two wheel steering with the back wheel reflecting a percentage of the front?
What is the theoretical top speed using hoverboard motors?
Ive just a question. Is that stuff u do is related to mechanics mechatronics or what section of mechanics
Mechatronics is a pretty good description once in involves things like servos to steer the rear wheel.
Did you try or use PID? The back wheel lags a bit behind
This is very interesting. Since you have the manufacturing capability, i'd actually be interested in a project where you made a sensible, practical e-bike, maybe suitable for off-road use, and with some handy electronically-controlled refinements that commercial models lack, like for example some sort of active steering or active suspension.
Cant you make 3d prints with better polymer or bake it in the over to strengthen it, or cover it in carbon fiber or fiberglass or something