Exhibitors & Products

Staffan Dahlström, CEO of HMS Industrial Networks, attended the flagship industrial show in Hannover 2019 where he gained an overview of an industrial landscape characterized by rapid change. “Five years ago we were the pioneers in innovative automation, the first suppliers of modern connectivity. Now the telecom and IT companies are making inroads into the market,” explains Dahlström. Above all, HMS Industrial Networks made a name for itself with Anybus. The engineers at Halmstad created a product family that facilitates network connectivity between machines or plant to middleware applications and more.

But is Staffan Dahlström concerned about these new competitors affecting business? “No, it is good that a new market is being developed. We are already cooperating with Nokia and Ericsson on 5G and are in the Proof of Concept stage with clients.” According to media reports, 5G is some way off ‒ it appears that TSN is a much more realistic option. “TSN is good, but last year we had more industrial Ethernet installations than Bus systems. It took a good ten years. We expect a similar development with TSN.” Dahlström is emphatic about the need to keep an eye on the installed base.

So is CAN a thing of the past?

I think it will take three or four years with the IT, and between ten to fifteen years for industry. But the fact is: TSN technology reduces the profit margin of HMS Industrial Networks and its competitors because the chips cost less. So CAN is obsolete?

“No, not at all. CAN is still needed. We can identify lots of new applications for this technology ‒ for example, in mobile machines or ships. This technology is becoming less important in motion control but it has in no sense become obsolete.” However, Dahlström and his colleagues need to remain diligent. The market is changing quickly.

Perhaps it is for this reason that Dahlström has been looking at a further, new area of business and acquisitions, including buying a new software company. HMS intends to put an Edge on the market so it can offer its customers direct machine learning (ML) on the machine. Dahlström says: “We collect data, we are responsible for its communication to a cloud application; we sort data, collect it from different machines and provide it with properties. So why should we leave learning on the machine to others?” He sees two options for data – Edge and the cloud. “Hybrid scenarios are the future. Our customers do not want a big data lake, but prefer to work with smaller data sets and the right data. For many users, the situation is a bit like not being able to see the woods for the trees ‒ the data are the trees, but they can’t see the woods, i.e. the applications.”

AI should not be left to the engineers

In future, HMS wants to transport data into the middleware via its gateways. From there, users will be able to push them into the AWS or Azure cloud and use them for training models. “But we can also train ML models on the Edge with small data sets and then put them directly on the machine.” HMS supplies connectivity hardware, the software and a Linux-based programming interface and API interfaces. “We are not the ML experts. We see ourselves as a partner for other companies with whom we work together and to whom we offer our gateways as a platform.”

Dahlström is convinced that there is no way to get around the Edge and eagerly offers automated driving as proof. “We can’t wait for cloud if a child jumps out onto the road. In industry we have the same latency problem with the cloud – slow connection. In the USA, for example, many factories have only just arrived at 4G level.” Furthermore, industry is conservative. “Not every company wants to transfer production data to the cloud,” stresses the company founder, spelling out his case: “With trustworthy AI, Europe can secure a lead position in the fields of artificial intelligence and machine learning – and challenge the dominance of the USA and China. It is not about ethical computing but about how companies handle data, models and algorithms and the policies they pursue. “AI and ML are great technologies, but we can’t just leave the issues up to engineers. We need control bodies,” says Dahlström.