Industrial automation trends from robotics to IoT sensors to machine vision are helping manufacturing facilities improve productivity and operational efficiency.
Today’s industrial automation trends are helping to make manufacturing facilities more efficient, productive and cost effective.
“Competitive necessity requires all manufacturers to re-evaluate their businesses, processes, supply chain and technology solutions to ensure their survival,” said Kevin Senator, CEO of
Bayshore Networks.
By adopting the latest trends such as collaborative robots (cobots), more IoT sensors, machine vision, etc., manufacturers can ensure they stay ahead and don’t fall behind.
Cobots — the new it robot
Robotics continues to be a major trend in industrial automation, with the popularity of cobots in particular seeing rapid growth.
Cobots are designed to work safely alongside humans without needing to be caged to protect workers, as well as increase the productivity of individuals on the production line. They also offer more flexibility than traditional robots, are cost effective and more easily deployed.
Based on research conducted by Interact Analysis, 23,000 cobots were shipped in 2019 with global cobot revenues exceeding US$650 million — representing growth of over 20 percent versus 2018. The company believes that cobot revenues are set for continued strong growth.
Material handling, assembly and pick-and-place are the three biggest industrial applications for cobots, according to Adrian Lloyd, CEO of
Interact Analysis. However, he also believes that cobots will grow rapidly in small manufacturing operations as managers realize that cobots, unlike traditional industrial robots, are highly affordable.
Deploying more IoT sensors
Industry 4.0 continues to push manufacturers to transform their facilities into smart factories. One way to do this is with IoT sensors, which have seen widespread adoption and is likely to have the biggest impact on manufacturers, according to Curtis Simpson, CISO of
Armis.
“Businesses are constantly seeking new ways to strengthen their bottom line, and IoT has provided a unique avenue to improve productivity while reducing overhead,” Simpson said.
While incorporating robotics into existing workflows might not make sense for a particular organization, Simpson pointed out that the use of sensors to collect data and better understand business performance can help nearly everyone.
IoT sensors are not only useful for improving operational efficiency, but also monitoring for issues before they become costly problems. Manufacturers of all sizes, large or small, can benefit from this type of insight into their industrial environments.
Simulation in manufacturing
Simulation in manufacturing refers to the use of software to make computer models to simulate manufacturing processes before doing real-life testing and to help at the product design stage.
Although simulation in manufacturing has achieved less notoriety than other topics, Lloyd noted that it is taking off in manufacturing. He attributes this growth to “the brutal market pressures that the sector faces — particularly the increasing need for manufacturing flexibility — which forces companies to redesign their manufacturing processes.”
To illustrate how simulation can benefit companies, Lloyd explained how an automotive startup in Thailand went from concept to first production in 21 months. By utilizing simulation software, the company was able to simulate the layout and operations of the factory before a single brick was laid.
“Such practices introduce radical efficiencies when laying out highly automated factories,” Lloyd said.
Predictive maintenance, cloud and more
Another interesting piece of the IIoT story right now is motor drives, also known as variable frequency drives (VFD), according to Lloyd. Motor drives control electric motors which, despite being a relatively simple piece of technology, are amongst the most critical pieces of infrastructure in any factory — without them, nothing can move.
“More users are now placing sensors and measurement devices in industrial motors to allow monitoring of status information for predictive maintenance. This points to a future where advanced software algorithms can predict motor failure before it happens. That’s where motor drives come in,” Lloyd explained.
Other trends include more adoption of hybrid on-premise cloud services, increasing the level of intelligence processed at the edge, and more. Manufacturers are also looking to have more control over their data and access to more reliable real-time feedback.
Machine vision improving operational efficiency
Machine vision overlaps with several important industrial automation trends (e.g., IoT, data analytics, robotics, etc.) all of which can rely on
high performance vision processing to improve operations. It can also help manufacturers lower end-to-end costs, deliver more consistent results and drive higher throughput, as well as increase customer satisfaction.
Luca Verre, CEO and Founder of
Prophesee believes enabling machines with more efficient, accurate and high-performance vision capabilities has significant potential to improve productivity, reliability and safety.
“Machine vision can be used to enable new levels of automation, freeing workers from tedious or dangerous tasks and improving overall efficiency of industrial operations, from manufacturing to assembly to packing and logistics,” Verre said.
Use of machine vision for quality control on production lines has led to direct and continuous improvements in terms of quality consistency, as well as significant reductions in downstream expense and brand loyalty challenges, according to Simpson.
“As production lines grow in complexity, this technology is also critical in ensuring that production lines can adapt autonomously to manufacturing variations, thus reducing downtime and increasing output and in turn, overall revenue potential,” Simpson said.
Applying machine vision to various applications
Machine vision technology has been deployed for many years in manufacturing applications such as pick-and-place and quality control. However, Lloyd noted that incremental improvements in terms of price-performance ratio of the products (e.g., increased resolution of cameras) and some big data techniques are taking the technology to the next level.
Today, machine vision is being used in high-speed counting of objects, preventive maintenance of equipment, spatter monitoring in welding applications, object tracking for logistics, security and safety applications, and more. In robotics, the need to be increasingly autonomous has pushed machine vision to become the “eyes” that enable them to maneuver successfully from point A to point B.
That does not mean machine vision is perfect. In certain applications where the machine vision system is deployed to assist quality control, Lloyd explained that users have found that computer algorithms have been inadequate at correctly identifying faulty components, leading to false positives or missed defects. By applying deep learning, a system can be trained to better identify faulty products with much greater accuracy.
“This significantly expands the scope and use of machine vision technology and will be an important driver for the market in coming years,” Lloyd said.
Machine vision also allows for virtualized interaction with physical assets, according to Senator, and could become a tremendous aid in training a new workforce as the retirements of older, skilled workers continue.
Next generation of machine vision
To develop even
more efficient machine vision, companies like Prophesee are using what is known as neuromorphic computing, which is essentially technology that mimics how the human brain and eye work.
For machine vision, Verre explained that it is represented by an underlying technique called event-based vision, which is an advancement over traditional frame-based approaches used for more than 100 years.
“Event-based vision captures only events that change in a scene, much like the human eye, rather than continuously processing very aspect of a scene,” Verre said. “This reduces the amount of data that needs to be processed in a machine vision system, thus improving performance and accuracy and reducing power consumption.”
Additionally, event-based vision is able to overcome the limitations of traditional vision where there is a tradeoff between data rates, speed and dynamic range. By addressing the limitations of traditional frame-based approaches, Verre highlighted this new machine vision generation enables a significant step toward new performance levels, more operating efficiency and excellence.