London’s financial district is hardly the place you’d expect to find a robotics start up. But, then again, the Capital’s start up tech scene is growing by the day.
Automata consists of two architects turned entrepreneurs that have decided to give their big idea a go. While it is easy to dismiss the buzzwords that include game changer, disruptive technology, and some kind of democratisation, after hearing about it, the cynicism does begin to wash away.
Frustrated by the choice of smaller scale and low cost robots in the market, the pair set about to fill the niche. There seems to be a stark contrast of availability; at one end is a £25,000 robot that costs as much again to install and maintain, or at the other is essentially a useless toy costing between £200 and £500.
“This was our inspiration and the launch point to design Eva,” said Mostafa ElSayed, co-founder of Automata. “It’s this idea of elegantly designed hardware meeting very simple software to become useful in ways that robotics, traditionally, hasn’t been able to in the past. We really believe this is a shift in what’s available.
“She’s a lightweight, easy-to-use, very low cost robotic arm. She is 2.2 kg, runs off USB or wireless protocols, and you can buy the hardware for just under £3000.”
Though the personal nature and personality they’ve given to Eva is slightly disconcerting, the pair are living and breathing this development. They claim to have made 13 hardware iterations in the last 16 weeks, with everything designed and made in its London offices.
What’s more incredible is that neither founder is a traditional roboticist... or even engineer. They have worked in the computational design research group at London architectural firm Zaha Hadid, and they are clearly more than proficient at coding.
Undeterred by lack of experience, the pair began by designing and then 3D printing the outer casing of the robot, sourcing and assembling all the electronics, and also writing from scratch the software to deliver what they say is a unique offering in the market.
“This is hardware development at the speed you’d expect from software development,” said Suryansh Chandra, the other co-founder of the company. “Most of our mechanisms are also 3D printed. If we design a part today, we can assemble it tomorrow.”
At full stretch Eva can pick up 750g. If the payload is closer, it can pick up 1 kg. Though making the mechanisms from 3D printed plastic means they are perhaps on the weaker side, for many of the repetitive applications the robot is aimed at, the payload is around this mark. And a big benefit is that it does not need lubrication. Though, again, this limits usage in one sense, in the other it opens up many more applications – including work in the food sector, an area still carried out largely by hand in the UK.
One problem that was encountered on early iterations was the motor overheating. However, the inclusion of fans means the problem was quickly solved. It was a simple case of ordering the fan components online, modelling the parts and connections in CAD, and then printing them out. The iteration was completed in days.
“We are aiming for around 3000 rotational hours in the life span of one robot,” said ElSayed. “For traditional users that is around 16 to 20 months of usage. By that point, there will be a new version of the robot with some more improvements.
“At this price point, we don’t believe it makes sense to sell the robot as a piece of hardware and say, ‘here, take it for £3000’. We want to sell a service and use a subscription model. You pay us £250 a month to cover your hardware, software, insurance, and if something breaks.”
At the heart of this product is software, which is key to delivering the intuitive experience non-robotic users will expect to easily get Eva up and running. To make it as simple as using a smartphone for the first time, the pair have put together software that makes the robot ‘learn’.
Though that might sound rather spooky, in practical terms, it’s a revelation. Users will teach the robot by example, literally by manhandling it through the motions and to the positions it needs to repeat.
“It’s called Teach by Example,” said Chandra. “But this kind of software is normally at a price point above £30,000.”
The pair basically wrote their own version of the software from scratch, enabling them now to sell it with the robotic hardware.
“The idea is that by combining very low cost hardware with incredibly easy to use software, you open up a whole new set of markets to robotics,” continued ElSayed. “Suddenly, for businesses where robotics just wasn’t on the agenda, Eva is the obvious answer.”
The applications are vast with a host of jobs that are repetitive and manual from fast food restaurants or coffee houses, to picking and placing factories, to start ups that want to upscale production and begin to industrialise a manual repetitive process. Many that might benefit from the use of robotics, don’t because of the initial investment.
“A traditional robotics company want a couple of months and tens of thousands of pounds to design a traditional robotic solution for you,” said Chandra.
Automata is currently working with a company making chipsets for Boeing, where the process basically involves dispersing epoxy. The company own a $30,000 deposition epoxy machine, but need a person to manually use a couple of syringes at one end. The big problem is the 20% error rate they are getting.
“He said to me, ‘I really don’t mind if your robot reduces my throughput, but just improve my error rate,’” said ElSayed.
Automata is currently looking for UK businesses willing to pay a subscription for Eva, and currently has more than 35 signed up. It is keen to begin running pilot projects to develop its roll out and begin finding new robotic applications.
“We thought manufacture had moved East, but in talking to a few companies, they are saying, ‘we’re moving back to the UK’,” said ElSayed. “These companies want to be running more efficiently, need to compete, and robotics is a big part of that. We’ve spoken to a variety of companies, but they all say the same thing: they’re small, can’t afford a big robot, but need to stay competitive.”
Chandra continued: “We’re not aiming to compete with the high-end robots such as Universal Robots or these kinds of companies. They prioritise precision. What we are arguing is there is a market for lower precision, lower payload use cases. And they are potentially greater than the traditional robotics market.
“What we prioritise is it’s easy to use and retask. I shouldn’t have to change the insurance, install secure fencing, or talk to a third party where they have to fly in someone from Germany to fix it.”
Automata is currently taking the robots to factories and piloting its roll out before it hopes to open it up to a larger audience later in the year.