Industry 4.0 is here, so let’s survey all the ways industrial automation is evolving with it. The promise of robots that can learn, share information, and adapt to changing demand is finally being realized.
There’s a lot of buzz these days around robotics, but manufacturing has relied on industrial automation for decades. Why is there such excitement now? The digitization of industry is occurring around a potent combination of robots, artificial intelligence, and related technologies. Still, any transformation involves both peril and opportunity for business.
From collaborative robots designed to work alongside humans to autonomous carts and drones checking inventory, there’s no shortage of innovations to follow.
Robots vs. jobs, or robots plus jobs?
First, let’s deal with the proverbial elephant in the room — robots and jobs. Ever since a 2013 Oxford University study titled “The Future of Employment: How Susceptible Are Jobs to Computerisation?” there have been widespread worries about how industrial automation would throw half of U.S. workers onto unemployment lines. At the same time, industry leaders complain about shortages of skilled and unskilled labor.
Tesla and SpaceX founder Elon Musk has publicly supported the Future of Life Institute, a foundation devoted to keeping AI “beneficial to humanity.” It hasn’t stopped cyberwarfare, accidents with autonomous vehicles, or the pace of industrial automation adoption.
In fact, even as countries worldwide buy and use more robots, employment (at least in the developed world) has not been harmed, according to the Association for the Advancement of Automation (A3) and the Organization for Economic Co-operation and Development (OECD).
Industrial automation is steadily increasing, without causing massive unemployment. Source: IFR
Of course, tasks and jobs will be affected by robots and AI, but the transition may not be as massive and uncontrollable as many fear. It is a challenge for policymakers and those who have to justify investments into industrial automation and re-engineer processes.
Enabling technologies for industrial automation
We can thank smartphones for the current robotics revolution. The GPUs, or graphics processing units, plus miniaturized cameras, accelerometers, and wireless technologies developed and tested on a massive scale for phones have allowed for more affordable robot sensors and processors for AI.
More specifically, advances in perception, object recognition, and machine learning allow robots to move safely in dynamic environments such as warehouses, manipulate a variety of objects, and share information with other robots and enterprise systems. The major chip makers, including ARM (now owned by SoftBank), Intel, Nvidia, and Qualcomm, are all betting heavily on growing demand from robotics and AI.
Materials science and biologically inspired designs have led to more dexterous grippers, such as those from Soft Robotics. For robots that aren’t yet fully autonomous, features such as teleoperation, augmented reality, and virtual reality are making it easier for one person to manage multiple machines at a time.
Cobots for all
A fast-growing area of industrial automation is collaborative robots, or cobots. Unlike traditional caged cells in a production line, these robot arms are slower, more sensitive to their surroundings, and easier to reprogram. As a result, they’re safer to operate around humans, although you should still review your production processes. Rethink Robotics and Universal Robots are leaders in this space.
Sawyer and Baxter are cobots from Rethink Robotics.
Cobots have also helped small and midsize enterprises (SMEs) get into the robotics game, as food processing and other verticals start to benefit from higher throughput and consistency. A wide range of end effectors offers swappable flexibility for small-batch manufacturing.
Note that there’s still an outlay, so companies expecting to immediately save money should carefully measure their returns on investment. A robotics as a service, or RaaS, model can help with both expenses and ongoing service.
What’s next for cobots? The mobile robots pioneered by Kiva Systems — which got snapped up by Amazon and created a vacuum that others are moving to fill — are now being paired with cobot arms.
Suppliers such as Fetch Robotics and IAM Robotics have developed supply chain robots that can go to a shelf, pick an object, and bring it to a human or another robot.
We’re just starting to see “lights-out” facilities, where human involvement is minimal. However, mobile robots are making the jump from being primarily in automaking to construction, healthcare, and retail.
Intelligence in the cloud
As with other forms of information technology, the hardware is challenging at first, but the software stack is the big differentiator. A common refrain at recent robotics conferences has been “Data is the new oil.” Getting the value of that data isn’t so simple, whether it’s crops needing more water or personalized footwear.
First, there’s the question of processing at the edge or the cloud. For heavy-duty big data, you’d want to send it to the (probably private) cloud. The bigger the data sets, the better it is for training with machine learning. This is why Google and Facebook have a lead in AI — although other companies are working to get more out of smaller data sets.
On the other hand, if you want a high degree of autonomy, for say, a humanoid or mobile robot, you can’t rely solely on Wi-Fi (even as researchers try to get around Moore’s Law with quantum computing). For now, a hybrid model is emerging, with mission-critical features like navigation and safety onboard at the edge and big data processing in the cloud.
IIoT connects robots, sensors, and other systems to the analytics needed to make data useful for humans.
The emerging Industrial Internet of Things (IIoT) takes data from all those sensors, feeding it to analytics and AI for translation into information that humans can use.
What can manufacturers do with all that data? In addition to being able to trace a product or component from order through production, packaging, shipping, and delivery, AI can help monitor shifting external demand and internal systems.
With predictive maintenance, you can know when a part of your production line is going to need maintenance and avoid disruption. In addition, simulation and digital twins can ease planning for automation or shifts in production to meet shifting or seasonal demand.
What’s next for manufacturing?
As agile, lean, and process automation speed along to the factory of the future, there are a few other things to keep in mind. The international economic environment is just as important as the machines and people inside your walls, and new uses for robotics and AI are still emerging.
China has been widely regarded as the world’s factory and the biggest buyer of industrial automation, but if it also becomes the biggest producer of robots and AI, how will that affect your sourcing? Trade and immigration policies will have a direct affect on manufacturing and consumption.
In addition, other countries and regions are specializing in parts of the global value chain. Manufacturing competitiveness depends increasingly on keeping up with AI and robotics.
Finally, adopting automation is not a simply a matter of plug and play. The most successful chief robotics officers, or those charged with implementing automation, have to be ready on multiple fronts. They must be able to justify robotics investments to cost-conscious superiors and replacement-fearing subordinates, manage the use of robots and software from multiple suppliers, and ensure that everything fits the business model.
Fortunately, each of these is getting easier as robots serve manufacturing in more ways.
Editor’s Note: This article also appeared in the May issue of Unmanned Systems, the journal of the Association for Unmanned Vehicle Systems International (AUVSI). It was distributed around Xponential 2018, for which we’ll have more coverage soon.
