Simply speaking, MOTOBOT is Yamaha’s autonomous motorcycle riding humanoid robot. But how and why have Yamaha decided to try and build a robot that... Rossi vs a Robot. WTF is Yamaha’s MOTOBOT?

Simply speaking, MOTOBOT is Yamaha’s autonomous motorcycle riding humanoid robot. But how and why have Yamaha decided to try and build a robot that can beat nine-time world champion Valentino Rossi round a track? And more importantly, have they achieved it?

Well according to the guys in charge of the three year MOTOBOT project, it has all been done in ‘the spirit of challenge’ and they were under no illusions just how difficult that challenge was going to be. The idea, and the thing that makes the project so unique, was to take a bog standard, unmodified motorcycle (in this case a Yamaha YZF-R1M), strap the robot to it and send it off down pitlane to beat Rossi. That differs from normal ‘automated vehicles’ as usually the automation is built into the vehicle, rather than just plonked on top of it.

Of course, the Yam did need a little bit of faffing with, in order to fix the bot’s botty in place, but to all intents and purposes, the R1M was just a bog-stock bike. It had normal controls, a normal dash, normal bodywork and it even had an ignition key.

They started the ambitious project in partnership with world-leading research institute SRI International with the aim of competing with a real world class riders lap times round a real track, but before they could even begin to do that, MOTOBOT had to learn the basics. In 2015, using levers, pulleys, cams and bits of string, they had the thing weaving in and out of cones at a very sedate pace indeed. But everyone (or everything) has to start somewhere, right? In the same year, they got the thing making its own decisions about a thousands of things a second doing 100km/h on the straight.

The aim for 2017 was 200km/h, which they quickly surpassed, and before long they were doing laps of the track against the clock. Their ultimate goal was to go ‘beyond human capabilities’ but you don’t have to watch many of its laps to realise they were a long way from that.

When the MOTOBOT team invited Rossi for a race, it’s was far from a close-fought thing. The former world champ lapped over 30 seconds faster that his robotic rival, with the Italian MotoGP ace posting a 1:25.740 lap, compared to a 1:57.501 for the bike-riding ‘bot.

In all fairness to Yamaha and the MOTOBOT lads, that’s comparing the thing to Valentino Rossi, one of the most successful bike racers ever. He is a rider that, even now, only a handful of people in the entire world can hold a candle to, never mind robots. And a 1:57 round a two mile circuit is probably an average trackday pace. You wouldn’t get many ‘slow group’ lads lapping witing 30 seconds of Rossi, so when you look at it like that it’s actually quite an impressive feat.

If somebody keeps throwing time and money at this kind of thing, then I’m pretty sure that one day they’d get a MOTOBOT version 10 that can post decent ‘fast group’ spec lap times, and maybe even compete with a few club level racers. But I don’t think it’s going to go any further than that, not in my life time, anyway. I know computers can process a fuck ton of information really, really quickly, but a machine can never have the same ‘feel’ as a chimp, never mind a MotoGP rider. And all of the scientists in the world couldn’t write a program or a computer algorithm that matches a MotoGP riders mind.

I love hearing about projects like this, and I admire not only their ambition, but how far they have got. Let’s not kid ourselves though, there might be plenty that robots and machines are really good at, but racing motorbikes isn’t one of them.

5 28 votes
Article Rating
Subscribe
Notify of
guest
5 Comments
Newest
Oldest Most Voted
Inline Feedbacks
View all comments
John
John
3 months ago

Yamaha may claim this is just about doing something for fun or because its a challenge but this thing will provide Yamaha with absolute gold in terms of bike development information. Being able to set some riding parameters and have the bike do the same thing over and over again and the engineers make marginal tweaks to it to make there bikes the best they can be.

Tyre manufacturers would also be clamouring over each other to get there hands on something like this for the same reasons.

Honda are also developing something similar.

Tech49
Tech49
3 months ago

I could see this being fun if it turned into a ‘battle bots’ kind of thing. Have them be remote controlled but equipped with clubs and saws. The Roman mob will be pleased.

Wee Al
Wee Al
3 months ago

It’s impressive, but (apart from to see if they can), what’s the point? I understand them doing these things for automated cars etc. As people will use automated passenger transport. But I can’t see people queueing up to be Chappie’s pillion….
Just seems like a hell of a lot of effort and money. For what seems like little in the way of actual useable technical advancement?

William Hare
William Hare
3 months ago
Reply to  Wee Al

Off the top of my head:

Automated delivery and or emergency vehicles. Especially in congested areas where a narrow nippy bike can have a big edge. (especially something with short trips carrying small amount of cargo like a takeaway)

Developing possible future lane keeping/hazard avoidance tech for bikes that do have riders. (they may never be fully automated, but having the bike spot hazards and intelligently apply the brakes etc. is probably going to be a thing. And unlike a car, rider posture and movement are a big part of it so would need extensive reproducible testing like this)

Gaining new insights to the whole rider-bike relationship for race development. (imagine you could set something like this thing doing endless laps under conditions you’d never put a real rider under, with a level of consistency no human could achieve. Maybe that leads to a real race edge they would never have discovered otherwise?)

Stright up crash testing with realistic rider behaviour. (As you can imagine it’s very hard to simulate some crashes without risking peoples lives or just using less realistic dead weights)

Brand prestige for Yamaha as a company. We’re here talking about it right? I doubt they’d do it purely for advertising, but that still a big cost they can write off. Same as the race teams. You get to do R&D and boost your image and reputation at the same time.

And lastly; just testing AI and robotics under new situations to learn what they do. Yamaha might be more interested in factory robots, but that wouldn’t necessarily make something like this a waste of their time. The robotics team could just learn new useful things for other robots from the challenge of this alone.

Martin Sanders
Martin Sanders
3 months ago

Yet.
I’ve been reading the updates on this for a few years. It’s quite astounding to see how far they have come in such a short time. When we need to worry is when the robots start developing new robots…