“Meet Bob who is joining the team today. Charlie and Theresa, you are in charge of the one-hour introduction. After that, he should be ready to do whatever you want him to.”
On the surface, a pretty standard workplace introduction.
However, in this case Bob is short for robot. He/she/it comes from one of the companies that are pushing back the boundaries of what robots are capable of.
Smart, flexible, easy to instruct/program, capable of working side-by-side with humans — not to mention cheap. The list describes just some of the areas where robotics is making progress by leaps and bounds.
The tempo is such that while Bob might today be a slightly surprising choice, he and his kin could be the natural first choice for your new colleague in factories and enterprises within a few years.
“Robotics’ great leap forward of the last years has taken big industrial robots out of their cages, shrinking them, and putting them next to their human colleagues. Better sensors and software advances, as well as new standards for movement speed and force, enable humans and robots to safely work side-by-side,” says Olivier Grenier-Lafond, sales and marketing coordinator at Robotiq.
“If we look at human space versus robot space as a Venn diagram, then these technologies allow the overlap to grow considerably, which improves cost justification and permits new process concepts,” Jaroslav Tyman general manager at Automation IG, says.
In other words, the current generation of industrial and enterprise robots could well be your next colleagues and co-workers. Let’s have a meet-and-greet, in the form of a robot award ceremony.
Welcome to the robot Oscars
Like any award ceremony, there are several strong candidates in each category. The winners will have some people shaking their heads in anger and dismay and others nodding in approval. The “show” is to illustrate just how far robotics has advanced and some of what the current generation of robots can do.
Without further ado, here are the winners.
Category: Ease of programming
Winner: Rethink Robotics, Baxter
Description: Forget manuals and programming languages. Many new robots give a nod to the industry’s early days. They can be programmed to do a task by moving their arms and ‘hands’ around. The robot ‘remembers’ the movement sequence and can be taught to perform new tasks in hours.
Category: Human mimicry
Winner: Yaskawa, MH24 – Motoman
Description: Robots for enterprise and industrial use are generally developed to take over tasks that were previously done by humans. The first step is often teaching robots to mimic how a human carries out a given task/set of tasks. A core part of the process is combining robotics with other technologies like 3D modeling, motion capture, and data analysis. The robot’s ability to carry out the task/tasks will then often be further refined through an iterative process.
Category: Dual manipulators, one core
Winner: ABB, YuMi
Description: Traditional robots have had their own version of the Zen koan “what is the sound of one hand clapping?”. For robots, it has been “how do you make two hands with each their brain clap?” For many years, robots were almost exclusive one-armed. Moving to a two-armed setup with one central control unit — or brain — allows robots to carry out a long list of new tasks.
Category: Independent and mobile
Winner: Aethon, TUG
Description: One of the biggest challenges for enterprise and industrial robots is how to be mobile and autonomous. In normal day-to-day situations, you don’t want to have to tell a robot what to do every second minute. You want it to be able to independently navigate busy, unpredictable environments, like factories or hospitals. Navigating and making your own decisions takes energy. Thanks to advances in battery technology, robots can do this without developing a co-dependent, clingy relationship with their charging stations.
Category: Most adaptable
Winner: Universal Robots, UR 3/5/10
Description: A single robot capable of working with machine foundries, laser-cutting systems, form presses, milling machines, and packing and palletizing units. Oh, and with assembly, installation, labeling, handling and quality control. And all that is just within the car industry. The robot can also work with car manufacturing, farming, electronics, metalworking, plastic production, in pharmaceutical companies, and on science and R&D projects. This is thanks to advances in flexible robot technology and the emergence of companies specializing in building plug-and-play effectors, sensors, and grippers.
Where do we go from here?
Looking through the list of winners gives a sense of just how far robots have come in recent years.
Without having to wait for acceptance speeches that thank family and god (in this case a group of engineers), it is time to look at where robotics will likely go from here.
For example, robots are now easily programmable, but they are still a long way from being able to work in the way that humans do. Part of our cognitive process could be described as performing “looks kind of like” tasks. Without the need for new instructions, we recognize tasks that look similar to things we have done before. We then use trial and error and minor adjustments in our attempt to carry out this new task.
Robots still struggle mightily with ‘”looks kind of like.”
“Improvements in vision systems, decision making, cognition, gesture recognition, increase mobility, and power units are all areas that are being explored by various companies. However, the current level of technology means that direct human-robot collaboration in many enterprise and industrial situations is still merely around the level of ‘they won’t hurt us’,” says Erik Nieves, former Yaskawa technology director.
Progress within all these areas is needed to reach what can be described as true collaboration between industrial robots and humans — where the two parties cooperate and work closely together to complete a task.
Judging on the history of the robotics industry, it looks perfectly placed to make those advances.
Robotics has always been a relatively small industry. Less than 250,000 industrial robots were sold worldwide in 2014. For comparison, Apple sold 74.5 million iPhones — just in the first quarter of 2015. This means that, to date, advances in related fields have been vital to keeping robotics R&D costs reasonable.
“Robotics has often ridden the coattails of other technology industries. It has happened with computers, which made chips for robots cheap and available. Then the same happened with sensors, where consumer electronics led the charge. The technologies we need for the next push in robotics coincide with what we are seeing become cheaper and more available now — AI, cloud computing, and battery power,” Nieves says.
Is it goodbye to humans?
The next developmental steps in robotics are likely to come in drip-like increments. It will lead to changes in how robots work — as well as where we meet them.
“With the advances comes an increase in the number of tasks where robots are a viable solution. Robots will be able to replace complex systems in custom machinery. There will be more robots everywhere for more applications, and they will be more likely to use tools/strategies to learn and adapt rather than control their environments,” says Tyman.
With specialization might come what I personally refer to as a “heavy metal moment.” Ask any metal fan, and they will tell you that heavy metal is not just heavy metal. It is a densely populated rainforest of sub-genres.
The same is likely to happen to the umbrella term “robots” which to covers everything from Boston Dynamics’ Big Dog to a Roomba.
While autonomy will increase, Nieves believes that robots will continue to learn best through human oversight and guidance.
“I am a big believer in supervised automation,” he says. “Robots will function autonomously until they encounter a situation, problem or task that they are unfamiliar with. They will then contact a human supervisor who can instruct the robot what to do. The great thing about cloud computing and robotics is that as soon as you have shown one robot what to do, your entire fleet of robots will all instantly know how to do the new task.”
One of the greatest points of contention about robots is whether they are going to take thousands — or even millions — of jobs out of the hands of workers.
The short answer is no. Robots will not replace all humans any time soon. The sophistication of the human mind, combined with our bi-manipulator setup with top of the range, highly pressure sensitive end-effectors make us a difficult competitor for robots. That will continue for a long time, while the mechanical brotherhood slowly improves.
“When asked about this, I often point out that it is not like robots are coming to take over the entire factory. 90% of tasks in manufacturing have not been automated because it has not been practical to do so with standard industrial robotics technology. It won’t be robot, robot, robot. It will be robot, human, robot, human, robot, human — all collaborating,” Grenier-Lafond says.
In the medium to long term, though, there can be little doubt that robots, capable of working 24/7 will outcompete humans and take over more and more jobs.
While this sounds like a problem, it is worth noticing that in many countries, robots will become a necessity before that happens.
Demographic data shows that many countries are facing a situation with a shrinking work force, compared to the number of elderly citizens. The rhetorical question becomes: how can factories and enterprises remain functional, let alone become more efficient, without advances in robot technology?
It is a question that is pertinent to not only industrial settings, but also to areas like health care and care for the elderly, where robots are also starting to prove invaluable.
Interested in learning more about the future of manufacturing? Join leading manufacturing experts at Singularity University’s Exponential Manufacturing summit.
Image credit: Shutterstock.com
