The announcement last month that Rethink Robotics was closing its doors came as a considerable surprise to the industry. The company’s collaborative robots, cobots as they are known, have been pioneers in the field and are a common sight in factories all over the world. Nevertheless, it seems that, amid growing competition, sales came up short and a last minute acquisition deal fell through.
But what are cobots exactly? Rethink’s fate aside, what sort of impression are they making? And what trends, if any, do they suggest within the evolving relationship between automation and manufacturing?
Cobots are a new phenomenon, a little over twenty years old. The original idea – that there should be a type of robot designed to assist, rather than replace, human workers in the carrying out of certain work – was developed by General Motors in the mid-1990s.
The concern at the time was that there were certain tasks in industrial manufacturing, such as repeatedly lifting up heavy car batteries, that were ergonomically stressful or dangerous. GM engineers therefore got together with robotics experts to set about devising what they at first called ‘intelligent assist devices’ – machines that could support and help move heavy loads.
The designers were pursuing this crucial difference from autonomous robots: that the new machines should have power and precision but no dangerously independent movement. For example, an intelligent hoist system was invented to bear the weight of a payload which, to the human worker steering its movement, would feel as if it were floating.
Much of the subsequent research and development focused on advancing safety technology: systems to scan for or otherwise sense human presences and respond with effective power-limiting feedback. The first ISO set of safety requirements specifically relating to cobot applications was published in 2016.
The feasibility of these safety technologies has broadened the range of cobotic applications, giving this type of robot the freedom to move independently around humans without the need to be caged off.
And now, just a few years down the line from the first trials and prototypes, various companies are rapidly expanding a whole new genre of automation. Universal Robots was founded specifically to create table-top cobots for small manufacturing businesses, hoping to reproduce in that sector some of the revolutionary impact originally made by full-size industrial robots.
And cobots have made a splash. A number of distinctive features, points of difference from traditional autonomous robots, has made them seem almost indispensable, or at least the occupants of a very useful niche in the market. Accessibility and versatility are their standout virtues. They are cheap, portable and easy to program (and, importantly, reprogram).
One of the most popular workbench cobots weighs a little over 10kg and costs about $35,000 USD. This could be about the half the price of a new industrial robot, though actually works out much less than that once the latter’s peripherals, safety apparatus and bespoke programming have been paid for.
It is about as easy to program as a digital video recorder (DVR), or even easier, and typically done through intuitive software. Programming a pick and place operation can be as simple as punching in a start and finish positions with numbers for rows and columns in between. Alternatively, thanks to machine learning, tasks can be taught by manual manipulation: an operator can move the arm through the desired track which the machine will then remember and repeat.
A cobot can be programmed in this way to do one task and then, with a switch of end-of-arm tooling, be reassigned to something else. Where the machine has the sensors and memory to recognise its new attachment, no reprogramming at all is required. It is plug and play technology.
This kind of robot excels at tasks that are repetitive, precise and relatively lightweight (payloads of, say, up to 10kg). Packaging, palletising, glueing, welding, polishing and machine tending all make ideal work.
In an automotive factory, a cobot might assist a human worker in the cumbersome job of fitting shock absorbers into wheel arches. Alternatively, a mobile version might move around a section of aircraft marking out where its human counterparts need to drill brackets. The things are called upon for strength, stamina or accuracy as necessary.
In complex assembly work, such as that of motors and drives, a human employee might share a workbench with a cobot and, with the machine’s assistance, put together a more or less complete system rather than just one part to pass on down the line.
Cobots not only take care of but excel at the parts of work that are less congenial to human beings while people flourish in the aspects of work less amenable to automation. The equation comes very close to vindicating the optimism of those who said that robots would not replace people in the workplace but free them up to work better, more creatively.
There seems little doubt that cobots are shaping up to be a major component of the automated factory, with small and medium-sized operations, in particular, feeling the immediate benefit. A recent report suggests that cobots’ current 3% share of all robot sales is likely to rise to 34% by 2025.
After all, these small, adaptable robots are not least a telling symptom of how industry is changing, breaking down obsolete distinctions and interlinking processes formerly considered in isolation or in parallel. As the principles of integration, harmonisation and connectedness continue to bring unprecedented levels of efficiency to industrial and other systems, the arrival of the cobots could not have been better timed.