June 27, 2017      

Although manufacturing has relied on robots for some time now, the biggest changes coming to industrial systems will not be in the arrival of new hardware. Instead, new robots will be distinguished by their artificial intelligence capability and connectedness with other robots.

Making robots “smarter” has already proven to be key to developing more useful and versatile industrial systems. For Industry 4.0 and the factory of the future to be realized, robots must be able to apply machine learning and communicate with one another.

Business Takeaways:

  • The biggest improvements in industrial robotics systems will come from increased connectivity and intelligence of robots
  • Systems will be able to be programmed through a single set of instructions, and the robots will figure out individual tasks
  • Machine learning will remove the need to program individual robots
  • Innovations in robotic grippers can expand the usage of industrial robot systems

A new era for industrial automation

Manufacturing will be a primary beneficiary of AI developments and the growth of the industrial Internet of Things (IoT).

“Flexibility is the new productivity,” said Dan Kara, robotics practice director at ABI Research. “Industrial manufacturing has evolved from the era of mass production with the goal of just putting out large numbers of products to [the more flexible system of] programmable automation. The future will bring intelligent automation, which is even more flexible and productive.”

Combining AI and industrial automation will result in significant changes in how robots are programmed.

“If you have a large line of machinery, you can say what your goal is, and using data from machines, it will sort itself out and figure out how to produce the item,” said Carl Vause, CEO of Soft Robotics.

One could even imagine a situation where industrial robots would automatically move around a factory to prepare for new tasks. Or entire fleets of robots transporting materials could automatically change their routes or destinations without the need to be programmed individually.

“The issue of indoor navigation has already been solved,” said Kara. As evidenced by Amazon’s warehouse robots, and a host of other warehouse robots, connected fleets of robots can navigate large indoor spaces with ease, he added.

Industrial systems are increasingly relying on robots to manufacture and transport goods.

Robots can easily navigate indoors, a feature that could add much versatility to manufacturing.

“When robots are all connected to a single network, you don’t have to program them separately to perform the same commands, share information with each other, be able to avoid obstacles, or recognize changes to layout,” said Daniel Theobald, chief innovation officer at Vecna Technologies.

Robot-to-robot conversations

Robots are becoming increasingly connected, both to one another in the context of industrial systems and to larger networks.

IDC research predicts that by 2020, 60% of robots will require cloud-based software. It also expects that, by 2020, “40% of commercial robots will become connected to a mesh of shared intelligence, resulting in 200% improvement in overall robotic operational efficiency.”

Among the industrial systems that could benefit from connected networks of robots, Kara cited systems for assembly, palletizing, food handling, warehousing, arc welding, and materials handling.

Robots are also increasingly able to learn, rather than just be programmed. Programming robots to be able to pick up items of varying size and shape used to take individual programming, a tedious task that still often left unsatisfying results.

Brown University has had success in machine learning by using a robot that can scan objects placed in front of it, noted Kara at the recent PTC LiveWorx event. The robot can then use the data about the shape, size, and orientation of the object to determine how to pick it up.

If this type of machine learning proves successful, robots could share data, enabling connected groups of robots to learn how to better perform tasks such as grasping and manipulation of irregular objects.

Finding solutions beyond machine learning

What if it was possible to simply find a solution for grasping and manipulation that didn’t require precision programming or machine learning? That’s what Soft Robotics is trying to accomplish through its unique robotic grippers.

“If we really want to be competitive in manufacturing, we have to figure out the issue of dexterity in robotics,” Vause said. His company uses air to expand or contract a gripper in a manner similar to a human hand, while the soft material negates the need for programming the robot on how to handle each object.

Soft robotics may prove to be the key for advancing industrial systems in the food industry.

Soft Robotics’ grippers can pick up irregular objects without the need for precision programming.

Vause also noted that industries like the fruit and vegetable packing industry are already experiencing labor shortages and that automating the process wouldn’t cause massive job losses.

“With this capability, you can not only handle [fruits and vegetables], but also move your workforce,” he said. “Nobody wants to work in a 33-degree vegetable-packing facility. It’s these labor shortages that’s driving price increases.”

Related content on RBR:

Looking forward to self-learning industrial systems

In the not-too-distant future, factories, warehouses, and other industrial facilities could include vast numbers of connected robots that can collectively learn how to perform any task imaginable.

“Anything manufactured, moved, or managed will likely be done through robotics,” stated Theobald.

But Vause imagined a slightly different future, where humans and robots are in charge of different tasks that favor the strengths of each. “You can harvest with humans and pack with robots, and you’ll have a significant ROI,” he said.

The ability of robots to learn and to share those lessons has yet to be proven on a large scale, but manufacturers and robotics suppliers are betting on the proactive possibilities of smarter and more connected robots.