There is a lot of hype about how the Industrial Internet will transform manufacturing. At its simplest, the vision comes to life at the intersection of connected sensors and machines and software-enabled intelligence. The combination promises to deliver what manufacturers care about most: greater productivity, lower costs, faster innovation cycles, world-class quality and more flexible and responsive supply chains.
Today’s factories are already filled with equipment — including robots — that monitor and collect all kinds of data. Manufacturers love to measure things: productivity, cycle time, pass/fail rates, the list goes on. Increasingly, there’s data coming from other sources as well, including unstructured content like social media, images, and customer support call transcripts.
What is needed are reliable and cost-effective ways to listen for that crystal clear note that resonates above the cacophony. I see three significant changes taking shape today that are heading us in the right direction:
1. The world’s not perfect — and robots have to be able to operate in spite of imperfection.
For the longest time, robots would only work if everything around them was fixed. You essentially designed the workspace around the robot — which works in some factories, as we have seen in the automotive industry. But it doesn’t work for the majority of tasks being performed in most factories. And it certainly doesn’t work in our homes. What we need are more general purpose robots that work in the environments we have.
2. Robots have to get smarter.
Robots over the years have not been known for being particularly smart. You program a robot to take a series of steps, and it performs those steps. You can build in if-then-else statements, but essentially, the robot is executing a program. Not a smart robot.
This is where we are seeing major breakthroughs and innovations. We now have better robots — robots that can work in environments that are imperfect. Automation of the physical task is innovating rapidly. But where we are also seeing rapid innovation is in the automation of the cognitive task, with advances in artificial intelligence — including machine learning and deep learning.
Where this gets really interesting is when you bring the two together — the new ways of automating the physical with the new ways of automating the cognitive.
3. Putting advanced analytics and big data to work for innovation in manufacturing requires that we replace “humans or robots” with “humans and robots.”
So, where do people come into the picture? In my mind, there are two major shifts taking place in the area of human-robot collaboration.
First, for the vast majority of robots traditionally, the way you got a robot to do something was to have an expert program it. As we think about democratizing robots, this isn’t going to work. There are not enough roboticists or computer scientists in the world to program all the robots we’ll need on all the tasks we want them to do. The robots coming on-line today are trained by someone showing them what to do. In the future, humans will tell them what to do, and eventually the robots will simply watch and learn — from humans and from one another.
Second, we need to embrace the value of agency — or free will — that only humans bring to the equation. The new frontier in manufacturing combines robots that can deliver real-time, sensor-driven telemetry data with software and data platforms that can aggregate structured and unstructured data. Organized, categorized, prioritized and presented, humans can then do what they do best: understand the information, interpret it and make smart decisions about how to improve processes and ways to drive continuous innovation.
When placed side-by-side in the work stream, robots and humans will finally be able to work together on problem-solving, process improvement and much more.
At the intersection of physical and cognitive automation — that’s where manufacturing’s age of innovation lives, and why there’s never been a more exciting time than now to be a part of it.
Originally published on GE Reports.