Bradley Smith explains MotoGP traction control

[davidboda46]

It's a new article series by Mat Oxley @ Motorsport magazine.
http://www.motorsportmagazine.com/opinion/motogp/how-motogp-traction-control-works

And this is the graph you should be looking at (it's down in the comment section). The one in the article is simplified and not correct.
https://uploads.disquscdn.com/images/feac5c5f47b36c150ec413340283b712278fd684c3460120547d4d98c1afb801.jpg

Cheers,

/David "Gonzo" Boda #46

[HornetMaX]

I see Hawk coming ...

[Hawk]

I see Hawk coming ...

Haha! I think we all now, or should know that todays MotoGP riders even with the recent TC changes are still not fully controlling their own bikes by a long way. 

I think everyone knows my opinions of scrapping TC all together and get back to having the rider in total control of their bikes.... Lets see who really has talent and those who don't so much. 

Hawk.

[davidboda46]

In theory I agree with you Hawk, but I think the development of the engines and tires makes it pretty much impossible without some sort of TC. To much power and grip. Not that it would be impossible to ride them without TC, but I think it would be impossible too do the same lap times and I think the races would be a lot different. Would be interesting to have a race without any ECU, but I think the riders insurance companies would have an issue with that

Side note: on the other hand I would have probably been very thankful if the bike I was riding a couple years ago at least had ABS. It would have saved me a trip into the wild and my only ever forward somersault, and of course, pain, lots of pain.   

Cheers,

/David "Gonzo" Boda #46 

[Stout Johnson]

In theory I agree with you Hawk, but I think the development of the engines and tires makes it pretty much impossible without some sort of TC.

+1
(Hawky is just a romantic irrationalist )

Side note: on the other hand I would have probably been very thankful if the bike I was riding a couple years ago at least had ABS.

In terms of ABS I am doing a Hawky and say that I am glad that there is no ABS...


Getting to the article itself. Pretty interesting read. I do have 2 remarks.

• Under 4 the author writes "Note how the lean angle stays steady as a result of the spike in wheelspin – the rider is busy dealing with the slide and momentarily stops adjusting lean angle."
  I would doubt that being busy controlling the wheelspin, keeps the rider from adjusting the lean angle. That does not make much sense. It would be much easier to reduce the lean and use the increased grip. The real reason is, that the rider uses the wheelspin to get around the corner faster --> power sliding, e.g. 0:29 and 1:27 in the video
  https://www.youtube.com/v/QljqcqhF1oA
  
• Very interesting that they use the throttle butterflies differently. But the response characteristics seem rather rough, don't they? At first I thought it was a plot with milliseconds, but in fact it is seconds(!). And the way the torque is being applied, seems very rough. For some tenth or so there is high torque, resulting in too much wheelspin, so it is reduced, the wheelspin reduces; then the software sees there is room for more torque, throttle butterflies open more again, resulting in too much torque, too much wheelspin and the whole circle starts new. This probably costs time and excessive tyre abrasion. I wonder why don't use some sort of exponential smoothing technique to smooth it out a bit more... 

[HornetMaX]

Hawk wants 1980s' bikes back to the future. Not that that would make the insurance companies any happier anyway

ABS is now mandatory here in France for bikes with more than 125cc. If you ask me, it's a good measure (actually I would make it mandatory for 125cc too, and even below that if economically possible).

For racing it's more debatable. To be honest I'm not even sure how much of a safety measure it would be: how many crashes are due to braking errors ?
Personally I'm not sure, I'd tend to think you have more crashes (and with more serious consequences) due to spinning the rear or losing the front in a turn.
I wouldn't be against, but I'm not strongly in favor either ... don't know ... maybe for rain conditions only ? Don't know ...

• Very interesting that they use the throttle butterflies differently. But the response characteristics seem rather rough, don't they? At first I thought it was a plot with milliseconds, but in fact it is seconds(!). And the way the torque is being applied, seems very rough. For some tenth or so there is high torque, resulting in too much wheelspin, so it is reduced, the wheelspin reduces; then the software sees there is room for more torque, throttle butterflies open more again, resulting in too much torque, too much wheelspin and the whole circle starts new. This probably costs time and excessive tyre abrasion. I wonder why don't use some sort of exponential smoothing technique to smooth it out a bit more... 

What you see plotted is the torque requested by the "sum" of the rider + TC (green line in plot 5): the torque that really comes out from the engine (and the one that comes out at the wheel) will be different (if anything, the tyre itself has some sort of response time).

At any rate, the usage of TC has permitted to save on tyre abrasion (compared to no TC and letting the rider handle this) so even the rough-looking graph does a much better job than no TC. The "roughness" of the graph may not be a problem in fact.


P.S.
I'm inches away from starting a kickstarter campaign for a time machine, so that we can send Hawk back where he belongs

[Stout Johnson]

I know, I know MaX  But looking at the graph, if you were the engineer/software guru from Magneti, what would you think? I certainly would not be happy with it. If the current system always "overshoots", either in terms of power outage or power cutting, then I would definitely try to smooth it out. Looking at it you can easily see that there is either way too much or way too few power delivered. I know that it is easier said than done, you cannot predict the exact desirable power deliveryand there are more aspects to it than which are plotted there, but still I think there is much room for improvement. I would certainly love to give it a try and try to programme it better 

Another thing that is worth mentioning. Even after the lean angle is reduced significantly, the wheel spin still does not reach a constant level or constantly drops. It remains alternates around a medium level. At first I wondered what the reason for this alternation is? Throttle and torque are almost perfectly constant and the tyre should have much increased grip at the lower lean angle. Imo, most likely the track surface is not smooth enough and does provide differing levels of grip. Also there might be little bumps which make the tyres compress and decompress and which account for differing levels of grip. At any rate, it is very very interesting and it would be super-interesting to have a similar plot from GPB.

Btw, thinking out aloud: It would be a cool thing if we had a chance in GPB to programme "our own" TC and WC formulas  maybe Piboso could implement that? And it would be a server-side option to either request all riders to use a standard TC or to be able to use his/her own.

[HornetMaX]

I know, I know MaX  But looking at the graph, if you were the engineer/software guru from Magneti, what would you think? I certainly would not be happy with it. If the current system always "overshoots", either in terms of power outage or power cutting, then I would definitely try to smooth it out. Looking at it you can easily see that there is either way too much or way too few power delivered. I know that it is easier said than done, you cannot predict the exact desirable power deliveryand there are more aspects to it than which are plotted there, but still I think there is much room for improvement. I would certainly love to give it a try and try to programme it better 

Welcome to the world of control of complex systems

People tend to think smoothing things out means better results, always. Well, that's not really the case.
Smoothing things out means the TC response will be slower, delayed. And a TC system (unless able to predict in an extremely precise manner the bike behaviour, which is very hard especially giving the varying conditions), must be fast. Hence the need for a "relatively simple" TC system that "overshoots" often but reacts quickly.

Looking at the graph I cannot conclude there's a lot of room for improvement. There may be room indeed, but it's not the "form" of the graph that can tell it.
Anyway thea rticle itself stated that last year's TCs were better, but this could also come from the fact they were using additional and/or sensors.

Another thing that is worth mentioning. Even after the lean angle is reduced significantly, the wheel spin still does not reach a constant level or constantly drops. It remains alternates around a medium level. At first I wondered what the reason for this alternation is? Throttle and torque are almost perfectly constant and the tyre should have much increased grip at the lower lean angle. Imo, most likely the track surface is not smooth enough and does provide differing levels of grip. Also there might be little bumps which make the tyres compress and decompress and which account for differing levels of grip. At any rate, it is very very interesting and it would be super-interesting to have a similar plot from GPB.

I don't get your reasoning: in the 2nd graph, the yellow line (actual slip) is below the red line (requested slip) so in fact the wheel spins less than requested.
During the turn the target spin is lower because (likely) the tyre also has a sideslip (so it can slip less longitudinally).
As soon as the bike is exiting the turn and lean angle is going down (picking up the bike), the target spin increases (and the TC is no longer doing anything).

Btw, thinking out aloud: It would be a cool thing if we had a chance in GPB to programme "our own" TC and WC formulas  maybe Piboso could implement that? And it would be a server-side option to either request all riders to use a standard TC or to be able to use his/her own.

Whooah man, you're asking for trouble. It's pretty complex stuff. You have engineers doing this 24/7.

[Stout Johnson]

Welcome to the world of control of complex systems

Thank you for welcoming the totally ignorant...

People tend to think smoothing things out means better results, always.

I am not people and I don't. All I am saying is that there obviously is room for improvement and I would love test my ideas. If you think that it cannot be improved, that is fine.

Another thing that is worth mentioning. Even after the lean angle is reduced significantly, the wheel spin still does not reach a constant level or constantly drops. It remains alternates around a medium level. At first I wondered what the reason for this alternation is? Throttle and torque are almost perfectly constant and the tyre should have much increased grip at the lower lean angle. Imo, most likely the track surface is not smooth enough and does provide differing levels of grip. Also there might be little bumps which make the tyres compress and decompress and which account for differing levels of grip. At any rate, it is very very interesting and it would be super-interesting to have a similar plot from GPB.

I don't get your reasoning: in the 2nd graph, the yellow line (actual slip) is below the red line (requested slip) so in fact the wheel spins less than requested.
During the turn the target spin is lower because (likely) the tyre also has a sideslip (so it can slip less longitudinally).
As soon as the bike is exiting the turn and lean angle is going down (picking up the bike), the target spin increases (and the TC is no longer doing anything).

You are talking about something completely different.

Btw, thinking out aloud: It would be a cool thing if we had a chance in GPB to programme "our own" TC and WC formulas  maybe Piboso could implement that? And it would be a server-side option to either request all riders to use a standard TC or to be able to use his/her own.

Whooah man, you're asking for trouble. It's pretty complex stuff. You have engineers doing this 24/7.

You are aware, that the current TC system in GPB was not developed by "engineers, doing this 24/7", right? 

[HornetMaX]

Welcome to the world of control of complex systems

Thank you for welcoming the totally ignorant...

Wohaa man. what's wrong ?!  Bad evening ?

I just pointed out that the comment "it's not smooth enough, there's margin to improve" sounds .. uh, very suspect. No need to go ballistic.

People tend to think smoothing things out means better results, always.

I am not people and I don't. All I am saying is that there obviously is room for improvement and I would love test my ideas. If you think that it cannot be improved, that is fine.

You may re-read what I wrote, especially the part "Looking at the graph I cannot conclude there's a lot of room for improvement. There may be room indeed, but it's not the "form" of the graph that can tell it." Of course you're free to think otherwise.

Another thing that is worth mentioning. Even after the lean angle is reduced significantly, the wheel spin still does not reach a constant level or constantly drops. It remains alternates around a medium level. At first I wondered what the reason for this alternation is? Throttle and torque are almost perfectly constant and the tyre should have much increased grip at the lower lean angle. Imo, most likely the track surface is not smooth enough and does provide differing levels of grip. Also there might be little bumps which make the tyres compress and decompress and which account for differing levels of grip. At any rate, it is very very interesting and it would be super-interesting to have a similar plot from GPB.

I don't get your reasoning: in the 2nd graph, the yellow line (actual slip) is below the red line (requested slip) so in fact the wheel spins less than requested.
During the turn the target spin is lower because (likely) the tyre also has a sideslip (so it can slip less longitudinally).
As soon as the bike is exiting the turn and lean angle is going down (picking up the bike), the target spin increases (and the TC is no longer doing anything).

You are talking about something completely different.

Then please explain:  reading your sentence for the 10th time, it still doesn't make sense.

You are aware, that the current TC system in GPB was not developed by "engineers, doing this 24/7", right? 

Absolutely.

At any rate, it is very very interesting and it would be super-interesting to have a similar plot from GPB.

It would take me 10min to provide you this (actually I think I already have it, commented out in the code).
But my willingness to provide you that just went down to absolute zero. Good night.

[davidboda46]

Wow, play nice guys. No need to get upset. 

Anyways, my wish for ABS comment was not referring to racing, rather an incident I had on a gravel road with a Suzuki Bandit. Went in too fast into a corner, instinctively grabbed a handfull of brake, locked the front and into the ditch I went. Bike's front wheel stopped in the far side of the ditch, i did not...

Cheers,

/David "Gonzo" Boda #46