How Ultra Reliable 5G Could Transform Manufacturing

Although Bosch says very careful manual inspections safeguard its high quality standards, they are time-intensive. An automated approach should improve speed and efficiency.

The new automated system “doesn't get tired, so we have actually a 100% correct rate” in the trial, René Bonse adds. However, he cautions it may not be quite so accurate when inspecting parts that require it to spot more subtle optical deviations. The trial uses image recognition software developed by HD Vision Systems, which is installed on the server in the factory’s IT room. René Bonse anticipates the same system could be applied to many other optical inspection use cases, such as checking the pins of the ECU, the labelling, or for scratches on the housing.

Autonomous mobile robots could also be used to ensure production lines have all the raw materials and parts they need. To do that, “you will, of course, need mobility,” notes René Bonse. “If you basically have a robotic arm on an AGV, you can run it very stably and thereby automate the feeding processes in the line. This can also be a big benefit in the future.”

The modular solution used in the trial can be adapted to different use cases. The 5G connectivity enables Bosch Rexroth to split the control application between a physical ctrlX CORE controller onboard and a server in the IT room running a purely software-based virtual version of the ctrlx CORE. “There are advantages in having the automation platform basically stretch over these two controllers,” explains Gunther May, head of technology and innovation, Business Unit Automation and Electrification at Bosch Rexroth. “They have access to the same data when you have this reliable 5G connection running over TSN … this opens a lot of freedom as you can actually build some of these applications in server-like structures, IT-like structures and get much more flexibility regarding performance needs. So, upgrading your machine means you just add another server or a bigger one.”

With a private 5G network and local-licensed spectrum, there is little danger of interference from other spectrum users. “For a robot especially, it's important to have a stable and really robust connection of the wireless technology and I see that on 5G,” says René Bonse. “Unlike Wi-Fi, which operates in shared unlicensed spectrum, a private 5G network can guarantee a specific level of throughput and latency”, adds Fatih Ulupinar of Qualcomm Technologies. He also stresses the importance of 5G’s TSN capabilities, which “means, you're able to do one microsecond or better time synchronisation, which is only introduced in 5G. It is not available in the earlier versions.”

Another key feature of 5G is its support for “gated scheduling mechanisms that allows you to provide a timeline as to when the packet should arrive at the destination,” adds Fatih Ulupinar. “So you can guarantee certain cycle times.” He also highlights the importance of the inherent security in a 5G network.

Using the coordinated multipoint transmission approach, the private 5G network can achieve “six-nines or even better reliability without incurring any latency costs through retries of the packets,” Fatih Ulupinar notes. The use of an on-site edge compute server also keeps the latency to a minimum. “So all these technologies come together to solve this use case, which wouldn't be possible with the current version of Wi-Fi or previous generation of [cellular networking] technologies,” he adds.

Support for mobility is important to enable robots and vehicles to be controlled and monitored wherever they are needed. “Quality issues, of course, can happen in manufacturing and you cannot plan for it, so you will not have a station or robot ready on every production line, so we want a mobile system for automating unplanned optical inspections, which we can move wherever we need it,” explains René Bonse. “You need flexibility. It is not cost efficient to have a robot for this purpose in every line, because it doesn’t happen a lot. Sharing the robot between lines increases utilisation.”

The trial participants also believe the 5G positioning technology they are testing could be valuable for manufacturers. During the trial, the technology has been able to pinpoint the position of the robot and the AGV to within 40 cm. Fatih Ulupinar says Qualcomm Technologies is learning a lot about how to refine this technology for use in a potentially-cluttered manufacturing environment. “We are already working on improved versions of the positioning algorithms,” he says, estimating that the accuracy could be improved to below 10 cm.

Employing 5G millimetre wave (mmWave) networks could also make a major difference. Using the “positioning technology in mmWave bands gets much better accuracy than with the 100 MHz available in the 3.7 GHz band in a factory environment,” Fatih Ulupinar explains.

René Bonse envisions that the 5G positioning technology will be used to move equipment, robots and vehicles into the vicinity of each other, before cameras and sensors take over to handle the very precise positioning required to connect one piece of equipment to another.

For Robert Bosch, the trial is yielding valuable insights into how automotive electronics manufacturing could become more flexible. “We really want to build up the knowledge and the network to prepare for the future,” says René Bonse. “This was the first private 5G network that we put up at our location. It was the first time for us to work with the distributed PLC environment and it was the first time for us to connect a robot over wireless. So, there were a lot of new learnings and benefits for us.”