How The Robot ASIMO Helped Build Honda's MotoGP Electronics Package

The major leap forward which Honda's MotoGP bike has taken in the past two seasons has come in two distinct areas, the chassis and the electronics. Much of the work of sorting out the RC212V's chassis was done during the 2010 season, when the factory tried out five different chassis variations and several different swingarms, before getting the bike right early in the 2011 season. Parallel to the chassis, Honda spent two years improving their MotoGP bike's electronics, after poaching two of Yamaha's key staff to work on HRC's electronics package. The resulting machine, in the hands of Casey Stoner, proved unbeatable throughout the 2011 season.

A photo posted on Twitter by Augusto Moreno de Carlos, editor of the Spanish magazine Motociclismo, provides an insight into how the development of technology can take interesting paths. The photo shows how ASIMO, the robot Honda has built as a technology demonstrator and R&D project, provided some of the crucial technology for HRC's MotoGP machine. The multidimensional inclinometer used by the RC212V to detect the attitude of the bike is a direct development of the system used by ASIMO to monitor the robot's balance as it walks and runs. The inclinometer, consisting of a collection of gyroscopes and accelerometers, provides information on how the position of the bike is changing: Is the bike banked over in a turn? Is the bike wheelying under power, or pitched forward on the brakes? How hard are the braking forces? How fast is the bike being tipped into the corner?

On the basis of this information, the electronics package on the RC212V can change the engine power characteristics to help the riders control the bike better. By sensing that the bike is braking hard - especially by combining brake pressure information with data from the inclinometer about the attitude the bike is in - the electronics can regulate the amount of engine braking to apply. By sensing that the bike is leaned hard over, power delivery can be made smoother to prevent the rear tire from breaking traction too harshly. By sensing that the bike is being stood up hard on corner exit, power delivery can be ramped up more quickly, allowing the bike to accelerate harder as the rider gets the bike onto the fat part of the tire.

Although entirely logical when viewed in hindsight, it is fascinating that the electronics required to monitor a MotoGP bike should be derived from a walking robot. Bipedal motion - walking on two legs - is a massively complex undertaking, requiring managing a constantly shifting center of gravity, as the technical manual Honda issued about ASIMO shows in some detail. Though the speeds involved are much lower - ASIMO's top speed is 9 km/h while running, a little over a brisk walk - the complexities and the required speed of data processing are broadly similar; the fact that the two wheels of a motorcycle are relatively rigidly fixed together mean that transitioning between physical positions is gradual. The fact that the human legs and torso which ASIMO is copying have multiple degrees of freedom in their movements means that the number of variables involved are greater, and change at a much greater rate.

While the achievements of Honda in building the RC212V's electronics management package are many, their package does not by any means give them an insuperable advantage. The electronics used by both Yamaha and Ducati are equally complex, with Yamaha revealing at Valencia that the electronics package uses predictive algorithms to adjust levels of control to accommodate tire wear and fuel consumption patterns as the laps tick off. The Yamaha's electronics package constantly monitors the response of the tire and bike against the behavior calculated using data from practice sessions. Electronics strategies are constantly changed to adapt to the feedback coming from the bike, and new strategies calculated for the following laps based on that feedback. Yamaha, like Honda, uses gyros and accelerometers to detect bike behavior and adapt to it: two years ago, Yamaha switched their anti-wheelie strategy from data coming from the suspension travel sensors to gyros registering bike pitch. That meant that the wheelie was being detected before the front wheel left the ground and the front forks were fully extended, and power could be cut earlier, but by less.

What both the data from Honda and Yamaha show is that limiting electronics on MotoGP bikes - as Carmelo Ezpeleta is set on doing for 2013 onwards - is not simple. Data on bike attitude from gyros and accelerometers has become increasingly important, as demonstrated by the marginal effect that banning the use of GPS data has had this season. Arguably, banning data from inclinometer packages would have a much greater impact on bike control than banning GPS ever did. With no data from accelerometers, wheelie control would be more difficult, and the factories would have to rely on supension data again. With no data from gyros, there would be limited information on how far the bike was being leaned over, making it more difficult to alter throttle response and the way that power feeds in based on the angle of the bike. More control would be handed back to the riders, and away from the electronics.

Given the freedom to program ECUs as they wish, electronics programmers would soon work their way around the problem. Though precise data on lean angle and acceleration might be missing, data collected through the data acquisition packages can be used to simulate bike attitude quite closely. Using just the data from the brakes, engine revs, throttle position, selected gear and gear ratios, the position of the bike can be plotted remarkably accurately. Using that data, programmers can take a very good guess at the attitude of the bike, and adjust throttle response and engine mapping as required. It won't be as accurate as using inclinometers, but it will be more than good enough.

Even physically enforcing a ban on inclinometer data could be very difficult. Anyone carrying a modern smartphone is carrying an accelerometer and a gyroscope, as the ability to switch display modes as you tilt and turn the phone will show. The size of the sensors required is already tiny: one commonly used triple-axis digital gyroscopic sensor measures just 4mm x 4mm x 0.9mm. Cost is also not an issue: mounted on a printed circuit board, the sensor can be purchased for under $50. As sensors get smaller, they become easier to hide, leaving the only option for controlling their use monitoring the data coming into the ECU, or imposing a spec ECU on the series and restricting the parameters available to the programmers.

Controlling the growth of electronics in motorcycle racing is not easy, and given the technology crossovers between racing and other areas - as shown by the use of technology from a walking robot on a MotoGP bike - certainly goes against the wishes of the factories. Finding a compromise which allows the factories to perform useful R&D while allowing spectacular racing is going to be hard. But given the dire nature of the racing during the fuel-starved 800cc era, it is also going to be absolutely necessary.

Back to top


I was going to say that I wonder where this development will lead us, but extrapolating the processing, and therefore predictive/reactive power of the software I think the logical conclusion is that ASIMO's descendants will be riding for HRC some day not so very far away. It has been said before perhaps partly in jest, but Honda's vision for the future surely includes a Honda bike with a Honda (lightbulb) rider.
The mind boggles.

One of my laments for F1 is the extent to which it has become an arena for remarkable development along paths which are not applicable to any car I'll ever own, with every part of the car compromized to facilitate aerodynamic downforce.

An electronics package that wants to know where it is on track via GPS isn't particularly useful on a road bike. Though I'm sure they can still learn general things about bike management electronics from such a system.

Nevertheless, I find systems which use inclinometers, gyros, accelerometers, and suspension sensors much more intriguing, because here we are closer to a system which would be useful riding down a road that the computer doesn't have any advance information about.

Similarly, a system which is trying to predict tire degradation over thirty laps is less generally useful than one which is better at adapting to what's happening *right now in this corner*.

At least even lessons from the GPS systems were more likely to yield improvements to my 2015 Honda CBR than Red Bull's wings were going to do for my 2015 Civic...

Just putting this out there: does the argument about motorcycle manufacturers needing MotoGP to develop their technology really hold much water any more?

Running a top MotoGP team is expensive. Most things can be simulated these days, and for Valentino Rossi's salary you could probably hire quite a few mechanical and software engineers at the top of their game, plus some solid development riders. For the rest of the team's budget, and the cost of developing the MotoGP machines, you could probably do quite a bit of R&D. Whole new commercial airliners like the A380 are designed and built using CAD, simulators and wind tunnels these days, whereas design of a motorcycle is a much simpler and less critical application in terms of the cost of failure.

Manufacturers are primarily in MotoGP to showcase their brand and sell road bikes off the back of it by winning races. Sponsors are there to slap their brands on the bikes. When the factories graft a bit of new technology (like the Honda slickshift gearbox) on to a MotoGP bike that they also intend to use on the road, it simply allows them to brand them as bikes that 'utilise MotoGP technology'. But the overwhelming amount of development happens in the factory.

I'm not in favour of control ECUs or anything like that. Even if I don't entirely buy the 'MotoGP develops technology' argument, I think what WILL potentially discourage makers from taking part is taking away the points of differentiation that they can work on between themselves and the other manufacturers to try and find an advantage. Remember, ultimately the big factories want to WIN so their bikes are promoted and sold.

So in my opinion it won't be 'lack of development for road bike technology' that will scare the big makers away - as I've said, they can mostly do that with computers and wind tunnels. It's lack of bike-tweaking options to gain an advantage over their competitors that will do it.

Yeah, I always had my doubts about that statement. Sure, there has been a trickle-down effect in some areas. But looking at all those technologically advanced yet utterly useless supersports cluttering dealerships at the moment, I get the feeling that if resources spent on racing went directly into developing road-going bikes instead, we'd actually have much better road-going bikes as the result!

The real reason racing is useful to manufacturers is marketing, not R&D. Competition for user's dollar has proven itself to be a sufficient force to drive development in other fields - take computers for example. I don't see why motorcycles should be an exception.

The development argument is still valid, but not the version the MSMA like to peddle. The MSMA insist that MotoGP technologies trickle down to road bikes. In most cases, MotoGP systems have no application in the production market; however, the manufacturing techniques are important. Rapid prototyping, new materials, extreme simulation, complicated electronics (you know the spiel). If the MSMA can learn advanced calc in MotoGP, production-market algebra is second-nature.

As much as I like to slam the manufacturers for their communications and marketing ineptitude, their dilemma is clear. If a company like Suzuki spends $20M to develop state of the art manufacturing technologies, but the average punter thinks Suzuki is garbage due to their finishing position, why would Suzuki stay in MotoGP? Furthermore, who is responsible for communicating to the punters? Dorna or Suzuki? I'm not sure it really matters b/c Ezpeleta doesn't look like he cares, and Japanese corporations frequently do things under a veil of secrecy so people don't "see how the sausage is made". Even when the MSMA are doing something good, they generally don't trust the public enough to give basic details. Congrats to HRC for this press release. Not great, but at least they are trying.

Back to the point, if the manufacturers are developing new technologies with 8-digit-price-tags, but the punters have no idea what's going on, the only way to align the engineers with the punters is to write rules like the 21L fuel restriction. In the case of 21L, complicated manufacturing techniques and expensive materials both produce an on track advantage. When a manufacturer wins, the bike usually is designated as "the best" by the general public, and the manufacturer gets a bit of brand equity for their GP investment.

Unfortunately, the MSMA's game causes cost overrun. The MSMA rules also put pressure on revenues, which are derived from entertainment value. If no infotainment is happening, entertainment value is necessarily derived by the value of the spectacle, commonly associated with the closeness of the racing. Social problems are prevalent in the MSMA arrangement as well. The participants, except the celebrity riders and engineers, become dispensable b/c they cut into the development budget. Revolt is imminent. IRTA are willing to play the part.

Expensive development, and bad rules. Bob's-your-uncle, the entire racing world is copying NASCAR. Spec equipment everywhere, and ginned-up faux-competition.

Motorcycle racing fans should never embrace NASCAR; however, I think we should embrace a bit of pragmatism. If you put a 21-year-old kid on a 250hp bike, and, for your entertainment pleasure, you ask him to ride around at 200mph in a few mm of kangaroo skin, he should be entitled to a relatively competitive bike. Superbike operates according to that principle. I think MotoGP would be wise to move that direction as well.

So let me tie this all together, a spec-ECU might be a bit too far b/c of the development value of MotoGP, but strict horsepower controls are probably a good idea. Strict horsepower limits have killed car racing, but hp limits don't have the same effect on motorcycles. Motorcycles pitch and roll. That's why we love them (and hate them as we fly over the high side), and that's why MotoGP won't die if they ditch the fuel limit for a rev limit in 2013. Hopefully, the MSMA feel the same way.

Casey Stoner has said that TC should not be used to make the bike faster, but only as a safety net to protect the rider from injury (via a massive highside) if he makes a mistake. He also says that he was eventually able to persuade his technician to set up his RCV that way.

So, one wonders where the game really is: just how much of this stuff is Stoner really using, how much does he know he's using? Did the programmers just turn down the intervention until he was no longer aware of it, while telling him it was off?

As for the road application... to some extent it's insanity (and vanity): building bikes that have way too much power to ever be used on the road (how many 1000cc road bikes ever have their slow-action throttles fully opened at any rpm, let alone near max power rpm?), then adding a bunch of electronics to help the rider avoid accidentally using the power to kill himself. If a 750 was a superbike in 1996 and a current 600 has more power power, why aren't we riding and racing 500's?

You wrote: "How much of this stuff is Stoner really using, how much does he know he's using?" It really doesn't matter. He is always using by far less electronics than Rossi, or any other rider. They should look up to him!

A little reminder and perhaps an answer to your question:

Settling An Argument: Casey Stoner Talks Electronics And Sliding The Front At Phillip Island There's a lot of talk going around that you switch the electronics off at Philip Island. You always use less electronics than most of the other riders, so can you explain exactly what did happen?

Casey Stoner: No, we definitely didn't switch them off. I mean, if I switched them off, my electronics guy would have an absolute heart attack! He wouldn't let me do it. So, no, we didn't switch them off, but we always have them at quite a low level, anyway, so we just turned them down. In the race I really had to take it easy, because we knew that tire consumption and fuel consumption is an issue, so I wasn't able to really push as much as I wanted to. We knew we had some reserve anyway, you know. We spent that whole race just being, just preserving basically, preserving tires, make sure I didn't make any mistakes, and preserving fuel.

[ It really doesn't matter. He is always using by far less electronics than Rossi, or any other rider. They should look up to him!]

I think there are more useful ways to regard your competition than fawning adoration. Leave that for the fan(atics).

I don't. His example (Stoner) was totally wrong. Everybody who follows MotoGP knows that Stoner allways uses less electronics than most of the other riders. I'm not fan(atic), but I hate when people downplay somebody.

Whilst I find myself increasingly at odds with your opinions on a certain rider, you keep banging out great stuff like this article.

Truly astounding the electronic complexity involved. It's a wonder (and yet somehow reassuring) Stoner, Pedrosa and Simoncelli still managed to highside themselves last year. Ultimately it's still about the rider.

Good point Graham about how much does Stoner (or any of them for that matter) realise the trickery is still working for him / them in the background. I was thinking the same reading the report.

In this whole off season discussion about the direction of MotoGP and the stratospheric costs of current bikes I had Asimo in the back of my head when thinking of Honda. It will be a fine balance to get or retain healthy participation in the series and preserve it as the pinnacle of moto tech. At one end of the extreme is Asimo grafted into a bike churning out laps alone at Honda test tracks like a cruise missile with no rider, no competition, no sponsors, and no spectators. At the other end would be NASCAR.

Not a Honda Fan? I think there are more useful ways to regard your competition than spitting on company who has done so much for MotoSport. Leave that for the fan(atics).

Bike in garage? Honda. Brand of the four bikes owned before that? All Honda. Baseball cap I'm coincidentally wearing as I type? It's black with red wings and the word Honda below it. Count me a fan. Could even call me a fanatic of the brand. I'll take that.

My Asimo comment was trying to point out with humor that if the unchecked technology and money war among the big manufacturers continued we could look forward to the pinnacle of motorcycle racing becoming a closed door Honda R&D session. Of course that won't be the case but hey, their $750k gear box from 2011 costs most of what an entire CRT bike will be in 2012.

This will create a firestorm and I don't wish to demean the skills of the current MOTOGP pilots but they've got to realize these type of developments, if not stopped or severely restricted could result in the bikes going round the tracks with nothing more than Stoner or Rossi "dolls" strapped onto them. It's no longer a sport if it becomes little more than expensive and noisy R/C racing. Fans want to see the SKILLS of the riders more than they want to see the wizardry of computer programmers, as you can do that sitting at home in your underwear with a video game console.

You (and others) say that but I can still count the number of potential gran prix winners on any given weekend on one hand.

You're worried about the bikes becoming too easy to ride but only the riders who we know have the most talent are able to go the quickest with a lot of others never really coming close.

Did Honda significantly improve the bike or was it just a case of getting the right rider? DP's performance is the best indicator, two wins per season in 2006,7,8 and 9 , four wins in 2010 and three in 2011.

Credit is given to Asimo for the hardware, but more likely it's due to the Yamaha software guys. The (electronic)hardware on these bikes is nothing revolutionary, it has all been done in the aviation industry for decades.

Not much fact in this article, but lots of speculation.

As far as the 2011 world championship, history has already proven that it was the right rider that sealed that for Honda. However from mid 2010 that bike was gouging forward in development. On the hardware side Nakamoto explained that it wasn't one specific thing that radically changed but many small things that made the whole machine work better overall.

I agree that software probably had a sizable influence on making the RC212 the bike to be on in 2011. The hardware is not breaking new grounds technically even if Honda built their own gyro/accelerometer unit. It's the cobbling of all those components together into a system and the experience and talent of the HRC electronics boffins that make it come alive.

The Asimo angle is good though. HRC probably have very nice resources to borrow R&D from other areas of Honda. If they can build a walking robot that does back flips it's not hard to imagine the racing division calling up the robotics research guys to have a friendly lunchtime discussion.

re: "The (electronic)hardware on these bikes is nothing revolutionary, it has all been done in the aviation industry for decades."

yup, and "en masse" on car-side. we can thank the omni-present auto industry for the low price point we enjoy on all these sensors, hardware and software. inherently poor economics has always dicated moto-side as the last to see anything. if not for ever tightening emissions, the majority would still be draining float bowls as some manufacturers are still having us do despite the fact we're well into the 21st century.

Just when did mass market cars 'discover' the world of DOHC? If I recall correctly Toyota used to seek engine design assistance from Yamaha.

re: "Just when did mass market cars 'discover' the world of DOHC?If I recall correctly Toyota used to seek engine design assistance from Yamaha."

as did blue oval (ie. Ford), but not sure i follow...? what do double overhead cams have to do with the economies of scale the auto industry has created for sensors and electronics...?

My early 1970s four door FIAT had a DOHC motor that was originally designed in the late 50s. I think Lancia used the same basic motor.
All far earlier than the Japanese adoption of DOHC for mass market cars.

Kudos for Carmelo Ezpeleta trying to change things. I've had enough of computer racing. Let's remove the sensors, limit electronics to emergency TC (high sides), and implement that control ECU. The mfr's will be forced to make engine development, chassis development, and suspension development. They want R&D, we want racing. If a happy medium cannot be reached then the series will go away.

I wish the control ECU was implemented this year!

The other thing this underlines is the farce of the lease fees. The hardware is a couple of hundred $. The reproduction cost of the software is zero, the expense was only in writing it the first time, which they had to do for the factory bikes. By adding it into the lease fees for the satellites, they are asking the poor to subsidize the rich.

Or it can also be viewed as the factories milking Dorna for their sponsorship money as Dorna did partially sponsored satellite teams in previous years. This maybe the reason Carmelo Ezpeleta is pushing extra hard for as much CRT as possible in 2013, giving as little Dorna's money as he could to the factories, via satellite teams. This maybe a super compromised on Dorna when they asked the factories to increase the number of satellite bikes available but so far, only Honda and Ducati were able to do so.

It looks as if it's Microsoft writing the software and charging every motorcycle for a copy of the license to utilise it.

That's why I am reading U Dave. I always thought this is one way future mc's will go. Asimo is by most learnt people considered the most advanced android till now, in terms of humanoid motion management. It would definitely not be Honda anymore, if they created this robotoid just for prestige. I believe Asimo can move much-much faster, but for the need to present a non-aggressive image to people, when they see him. While its body's multiple degrees of freedom require lots of computing power, this is nowhere near the limit of current processors.
Where does all this lead to? My opinion is that a self-righting motorcycle is already feasible, the hindering factor probably being its 'introduction strategy' to the everyday rider: that is to a notoriously conservative guy --not to mention the fanatics who will (whenever something like this happens) declare Holly War on electronics and the molestation of the Sacred Interaction between rider and bike.
Thumbs down to all these, says I. I fully expect this fascinating new motorcycle to materialize during my lifetime (ain't extremely young anymore, mind U), as I wouldn't mind a bike that could save my neck next time out. Thank you David.

Just regarding processing power there:
The human brain can guide movement at a pretty damn reasonable level of reliability, and it actually runs a kind of simulation of body movement ahead of time to give you options. This is something you can sometimes experience; I was fielding in a game of cricket, and a ball was struck that I could catch - but it was going quickly and I had to react. I know I thought about sliding on my knees to it - right height for the catch but there was no way I'd ever accelerate quickly enough to get there. Discard option. I don't remember anything after that until I was horizontal, airborne, and holding the ball (which is the mind devoting processing power to my actions and suspending memory - it's actually quite common). The period I do remember was like the common comment in sport 'It all seemed to happen in slow motion...'.

This actually can be extended to riding a bike/driving a car. One that is very skilled will 'feel it is a part of him' - it's like an extension of the same process going on in the brain. The ability to do this using a digital brain is currently beyond not just processing power, but to some degree architectural, hence the research into computers that learn by mimicking the biological learning process.

Whilst no one fully understands the machinations of the mind, my opinion comes about from an excellent lecturer at my uni. He was a programmer, but had taken a keen interest in Artificial Intelligence and the subconscious mind. One part of his interest was in comparing a biological computer to a silicon one - it's not entirely apples and apples - but there are ways of guessing. One method would be to take a fairly understood group of nerves, like the eye. The eye does a lot of functions which computers also do - line enhancement, movement detection etc, so it is possible to make a rough extrapolation to the computing power of the brain. The confounding part is that this method falls short in a few areas. The other problem is that the brain consumes ~40W and makes next to no heat. If you get a high performance desktop computer that can't match the brain, and look at the TDP of the CPU (up to 125W) and the power consumption for the overall computer (I have a 600W peak PSU), then we have a massive efficiency problem as well.

Therefore, you can't say that ASIMO is actually near being able to fully exploit the multiple degree of freedom movements of his hardware to the degree that a human is. And therefore, I don't think you can say that a self-righting motorcycle is feasible in terms of computing power - especially in a race or even traffic situation - it isn't that dissimilar an operation to judging how to catch the ball.

In my opinion, of course.

A few thoughts:

-Robot 2 wheel vehicles exist and work. There is even a humanoid robot that can ride a bicycle. It's just a matter of "better", not "is it possible?"

-Look-ahead simulation is exactly how chess computers work. It's a long time since a human has beaten the best machine. No reason it shouldn't be feasible for physical simulation. Even better, it's eminently parallelizable, so it gets around the current sticking point on scaling and speed.

-Wasting a couple of hundred watts on electronics isn't an issue with a bike generating nearly 200kW. The high pressure fuel pump probably consumes as much.

Know what technology I'd really like to see? How about that look ahead simulation in traffic networked to other vehicles? Have it intervene when you're riding down the road and that person who's pulling out from a side street without looking. Bike bike gets programming flag that it has right of way. Car gets flag that it does not. Car driver steps on accelerator without proper checking of clear road. Bike computes closing distance to car and sends "right of way" command to car. Car interrupts ignition leaving no power under accelerator until bike passes. Now that's some technology I could get behind!

you know that any such system would:
a) impose all speed-limits by force;
b) should you manage to avoid a), provide detailed records to law enforcement agencies post-crash of easily prosecutable "explanations" such as excess speed...

"The multidimensional inclinometer used by the RC212V to detect the attitude of the bike is a direct development of the system used by ASIMO to monitor the robot's balance as it walks and runs."
Would not this part of the electronics package be very transferable to road bikes, and according to the article is not expensive?