Robot Scrum Machine

[lincoln]

I like Tom's wikipedia page.

The little brother of Ed Carter, he's seen to many as the least talented of the Carter boys.

And Nick Farr-Jones was less talented than his brother Peter - something called true grit comes to the for every time.

 

[RugbyFuture]

scrum simulator they invented in conjunction with Thales, the company recently awarded a $1.78 billion contract to develop Europe's next generation of weather satellites.
http://www.theaustralian.com.au/news/sport/rage-against-french-machine/story-fn6atwd0-1225960476969

arent they the guys who make some of the weapons for the ADF?

 

[Groucho]

arent they the guys who make some of the weapons for the ADF?

They do indeed make systems for ADF, not just ordinance but clever stuff like intel systems and avionics. Coincidentally, they are also one of the companies my company is partnering with on remote automation systems for oilfields.

 

[matty_k]

The fact that it can be programmed with scrummaging tactics from various nations is the cool part.

 

[dougdew99]

I assume, now, that all you guys who said it was a pipe dream are eating your words, and recognising my status as a Rugby visonary... too bad the French got in first... come on Aussies, let's see the True Blue version ASAP

 

[matty_k]

come on Aussies, let's see the True Blue version ASAP

What? It claims that it opened a dodgy email and doesn't come to work on Monday.

 

[Scarfman]

Wow! Bruce, you have to step up here. Get onto TPN when he gets back and get things going.

Save Australian rugby, Bruce!

 

[Crow]

I find these quotes interesting:

But given France was able to build its prototype for about $275,000 -- the cost of a fringe Wallabies' contract -- the ARU could hardly spend its money to better effect than by tapping into this technology.

and

Totally unskilled players were able to generate 800kg of force. A scrum comprising under-20 players recorded in the 900-1200kg range, while the French Test pack smashed into the simulator with 1400kg of force.
Given that an opposing pack would come in just as hard, something approaching 2800kg of force is unleashed for 80 milliseconds at each engagement.

Alarmed by its discoveries, the FFR immediately banned impact on engagement in community rugby scrums. If for no other reason than player safety, the ARU should take note of France's findings and response.

Given the improvement in understanding the forces involved, I'd be surprised if the IRU didn't move on scrums post world-cup.

 

[Rob42]

Wow! Bruce, you have to step up here. Get onto TPN when he gets back and get things going.

Save Australian rugby, Bruce!

Sounds like Groucho has the inside running on the software side, so just give him, Bruce and TPN 100K each and we're all set. But honestly, is there any chance they really built this for 275K? Even if it's all based on some other platform, eg a full-motion flight sim, that doesn't sound like much.

 

[Groucho]

Thales tackles scrum simulation
Neuilly sur Seine, 4 June 2010 – Thales announced today the inauguration of its rugby
scrum simulator— the first of its kind in the world — developed in close partnership with the
Fédération Française de Rugby (FFR).
The simulator was designed to analyse the risk of scrum accidents, particularly spinal
injuries, and thus improve player safety. Instability and collapse of the scrum formation is one
of the primary causes of player injuries. This simulator meets critical requirements for
reducing accidents, but also goes one step further, as a tool for coaching purposes and
match preparation.
Conducted as part of a management-transition training programme at Thales, the project has
combined state-of-the-art technology and advanced research for the direct benefit of toplevel
sport. A group of seven Thales personnel worked closely with the teams at CNRS and
the FFR for several months. Development was led by Didier Retière, in charge of the French
forwards, Pierre-Paul Vidal, CNRS research director at the Université Paris Descartes and a
specialist in neurosciences, Julien Piscione, senior research consultant in biomechanics and
head of the FFR’s science unit, and Serge Couvet, simulation engineer at Thales.
The scrum simulator takes the shape of a six-legged robot, relying on a six-axis motion
system to respond to player inputs (force and motion) via sensors installed behind the
beams/shoulder pads.
The simulator reconstructs the scrum situation by moving the beam left and right, backwards
and forwards, up and down, combined with three-axis rotation. Unlike a simple muscle
development simulator, which measures forces only, the simulator is designed to develop
sensory-motor control. Individual player weaknesses reduce the overall effectiveness of the
scrum formation. The simulator identifies these weaknesses so as to making the scrum more
steady and stable.
“The scrum members need to make the formation move as a single man,” says Serge
Couvet, project engineer at Thales.
Sensors located between the beam and the simulator structure measure the engagement
forces of each player. The simulator – which is designed to support player qualifications of all
levels – reacts in real time, according to a pre-programmed control strategy.
“The simulator is a real revolution,” says Dr Julien Piscione, senior research consultant in
biomechanics and head of the FFR’s science unit. “Not only is it the first simulator of its kind
ever developed in the world, it is also the first time that neuroscience has been applied to
simulator design. The secret behind this innovation lies in its ability to generate
proprioceptive inputs. These allow players to decide how to move and push against the
simulator, which reacts accordingly.”

This mirrors the discussion we were having earlier, in which the machine progresses from a single plane device, to a multi-plane device, to (eventually) a multi-robot device.

This robot scrum is at stage one of the progression: it provides the fixed man/machine interface of a conventional scrumming machine, robotised to the extent that it can move in three dimensions and is reactive to player inputs. It is essentially a single haptic device, and given the size of the forces involved, its movements are probably quite rudimentary.

matty_k is right: the cool part is that it can be programmed with known scrumming strategies.

 

[Rob42]

Groucho, that really sounds fascinating. I stand corrected on my earlier cynicism. Sounds like it can register a lot of data about the humnas opposing it, as well as providing a simulation of an opposing scrum.

I wonder if the Italian program includes relentless boring in at 45 degrees?

 

[Groucho]

Groucho, that really sounds fascinating. I stand corrected on my earlier cynicism. Sounds like it can register a lot of data about the humnas opposing it, as well as providing a simulation of an opposing scrum.

I wonder if the Italian program includes relentless boring in at 45 degrees?

I was thinking about that, Rob42.

I guess that a fixed front row device like this one has wouldn't be able to respond meaningfully to movements that are intended to deform the structure of the front row, like boring in. I think for that you would need the machine front row to be made up of three separately hinged devices, i.e. a stage two device. Then, when a human prop bores, in the machine prop reacts to the movement, first by resisting, and then by failing at a certain pressure level, which has a de-powering effect on the machine scrum.

Of course boring in is cheating. No one would program a robotic device to reward cheating.

 

[Scarfman]

I reckon the robot scrum machine (RSM) should have heads and shoulders which move, rather than just soft straight pads.

 

[matty_k]

That would be ideal. Has there been a photo?

 

[Groucho]

I reckon the robot scrum machine (RSM) should have heads and shoulders which move, rather than just soft straight pads.

Yeah, I agree. That'd be especially important in a phase two device, in which the RFRs (robot front rowers) have independent movement. Perhaps the robot props would have fixed outer arms as well, to facilitate a realistic bind and apply a realistic force.

In a phase three device (where the RFRs are fully robotised) then those outer arms would move. But that is a significantly more complicated machine.

 

[Groucho]

That would be ideal. Has there been a photo?

 

[Scarfman]

I suppose Phase 4 is giving the robots artificial intelligence.

In Phase 5, they would be fully human and you could scrum against them.

Hang on ...

 

[Groucho]

 

[Groucho]

But honestly, is there any chance they really built this for 275K? Even if it's all based on some other platform, eg a full-motion flight sim, that doesn't sound like much.

That's true. They built this as an engineering exercise as part of a management course for their engineers, so the people cost will have been free and they would have had the use of Thales' considerable resources. That 275k would just have been for the technology, I think. Provided you can use parts off the shelf, the main cost is always going to be development manpower, which could double or treble that.

 

[Scarfman]

Crouch, touch, pause, engage!