RoboDog leads to another pet project

Its doggy antics and e-mail reading ability give the robotic canine an endearing novelty status.But for its designer its successful performance points the way to more serious applications.

Remember that RoboDog stunt last year? It was July. You were watching the 10 o’clock news. It was a time when the world seemed a safer place. You got to the end of the broadcast and the newsreader launched into the ‘And finally’ slot. ‘So is this the future of the four-legged friend?’ he asks chirpily. ‘A robotic dog that responds to voice commands – and will even read out your e-mail…’

RoboDog, about the size of a labrador, wowed TV audiences across the globe with its doggy antics – cocking a leg, sitting up to beg, even walking around carrying a small girl on its back. But for its maker, British technology company RoboScience, it was just a staging point on a route to a much more serious commercial project.

Nick Wirth, a 35-year-old Formula 1 design ace and the engineering brains behind RoboDog (designated the RS-01), is giving nothing away. All he will say is that the new project will be some kind of robotic device aimed at getting images or information from places that are impossible or unsafe for humans to access. It’s aimed at an industrial market, and also at the police and military. It will be simple to operate, and ‘hugely cheaper’ than the kind of equipment or systems that are available on the market today for this purpose.

But will it will have wheels, or tracks, or legs like the dog?

‘I’m not telling you,’ Wirth says. ‘But when you see it, it will be obvious.’

Flexible design set-up

RoboScience, owned by Wirth and business partner Mark Oates, is seeking funding of ‘a few million pounds’ for the project, which apparently is small beer compared with similar development projects in robotics.

It’s a low-budget, low-overhead operation, using five freelance engineers, more or less working from home, designing remotely and sharing files and data on Unigraphics software (now part of EDS PLM Solutions).

The six-man team (including Wirth) is the same line-up that created the original RS-01. Roy Potter on mechanical design, Mike Albery on electronics, programmer Simon Turner assisted by Pete Disdale, and Mark Tyson in charge of construction.This flexible set-up has also helped the project weather the current economic crisis, which has put the brakes on development work. According to the latest estimates, the new product might not be revealed before next summer at the earliest.

In the meantime, Wirth is set to announce his return to motorsport engineering within the next two weeks, working in parallel with his RoboScience commitments.His last job in this field was as chief designer at Benetton, a post he left in 1999 after becoming frustrated over a series of management issues.

Wirth is an engineer by discipline who loves nothing better than to design, rather than be an administrator. But he also has a strong entrepreneurial streak that has helped him carve out a whirlwind career in motorsport in the space of just 10 years.Wirth’s CV is jaw-dropping. Having picked up a first-class degree in mechanical engineering in 1986, he embarked on a career in motorsport with March Engineering – a job he gained partly on the strength of a summer job he did there before graduating.

Strong entrepreneurial streak

At 21, he was working at March with Adrian Newey (now technical director at McLaren) on Grand Prix car design. In October 1988, aged 22, Wirth was offered the job of chief designer by French racing team Ligier. He turned it down, but a year later, at the age of 23, he formed high-tech racing company Simtek with business partner Max Mosley (later president of the F1 governing body the FIA).

At 24, Wirth designed, in secret, a F1 car for BMW, though the project was shelved. At 27, he formed his own Grand Prix team, Simtek Grand Prix, and then watched as one of his drivers, Roland Ratzenberger, was tragically killed in a high speed crash at the San Marino Grand Prix in 1993.

Two years later the team, which had been struggling financially, was wound up with $9m debts. Wirth was on the job market and went to join Benetton’s F1 team. He became chief designer at the age of 32.

But after around three years at Benetton, Wirth quit. He found he was spending too much time embroiled in team politics, management and administration, and not enough doing what he liked best: designing.

Wirth turned to robotics to reconnect with his engineering passion. But the challenge he set out to face was a very specific one. Making a big, strong, but incredibly lightweight robot: a classic engineering challenge.

Sidebar: The making of RoboDog

According to Wirth ‘the problem with robots that have legs is that the laws of physics are against you when you start to make them bigger’.

He explains with a few jottings on the back of an envelope. ‘Imagine you have a quadruped of a certain size that has been made with the best technology around, in terms of batteries, servos, motors and so on. This gives it a certain density. But supposing, once it’s built, it’s not big enough, or it’s too slow. So you decide to make it twice as long. That’s when the trouble starts.’

Wirth explains the physics. If the robot’s dimensions are doubled, its volume increases by eight times (two cubed). If you use the same construction materials, then its mass will be eight times as big too. But for any given position the quadruped adopts, the torque on any joint will be 16 times as great as the original, because the mass is eight times bigger, and the lever arms are twice as long.

So to get the same motion speeds, even after doubling the size of the gears, motors and batteries, you would still need to use twice as much power, or settle for half the battery life.

‘What’s more,’ continues Wirth, ‘the load on the limbs will be eight times greater, but the cross-sectional area of the limbs will only be four times greater. That means the stresses on the limbs are double those of the smaller robot. So in effect, everything is working against you.’

It is this problem of mass, torque and stress that Wirth set out to remedy, using a combination of the packaging of the motors, gearbox and bearings at the joints into a tiny space as well as in the choice of materials and how they are used.

‘Anyone could pick up an engineering book and select the highest-spec materials in terms of strength and weight. But it’s how you use them that matters,’ he says.

Striking difference

On this basis, the difference between the Sony Aibo and RoboDog, which it preceeded by six months, is striking. ‘If the conventional construction technology of the Aibo [which is 25cm long and weighs around 1.5kg with batteries fitted] had been scaled up to the size of the RS-01, it would have weighed around 30kg,’ Wirth says.

By contrast the 82cm long RS-01 weighs 12kg.

The trick involves the extensive use of kevlar, magnesium, and carbon fibre, but also relies on the design of the servo joints, combining elements that are usually separated, and thus giving a much better power-to-weight ratio. At the same time, much of the versatility of the RS-01 has also been the result of the power electronics and battery technology that havebeen incorporated into the system. The RoboDog brain is in effect a miniature PC running Windows, with wireless networking to its owner’s PC, allowing it to be permanently on-line.

At its first public outing last summer, much was made of the fact that a small number of the RoboDogs would be manufactured for sale to the public, with a price tag of around £20,000 each.

This has yet to happen, though RoboScience says it is in talks with a consumer ‘white goods’ group (which is not Dyson), to license the technology, and provide manufacturing, distribution and maintenance services.

But don’t expect to see one in your local electrical showroom quite yet.