IoT news of the week for Nov. 4, 2022

Xerox is spinning out its predictive maintenance technology: Xerox’s Palo Alto Research Center (PARC) has spun out Novity, a predictive maintenance technology startup. Like other predictive maintenance software, Novity uses equipment sensors and proprietary algorithms to help manufacturers keep an eye on anomalous data in their plant operations that might cause future problems. A Xerox release says the Novity TruPrognostics engine uses both ML and traditional physics-based models of equipment, and can predict failures with 90% or better accuracy. Xerox will maintain a minority stake in Novity. Predictive maintenance is a popular rationale for installing IIoT sensors and there is tons of competition in the sector. But we’ve yet to see consolidation, so why not add one more option? (Xerox)

Startup pitching cloud-based PLCs gets $10M in funding: This is pretty noteworthy, mostly because the sheer audacity of selling a cloud-based programmable logic controller (PLC) is breathtaking. PLCs are highly specialized, rugged embedded computers designed to take in data from sensors on manufacturing equipment and operate them. Redundancy, low latency, security, and absolute rock-solid performance are essential, which is why most people I talk to think it will be half a decade before we even see those devices get wireless connectivity much less see a virtualized PLC that runs in a data center hundreds of miles away. But Boston-based Software Defined Automation has scored $10 million from investors to make cloud-based PLCs a reality. I can’t wait to learn more. (Software Defined Automation)

This safety vest uses TinyML to sense danger: Swedish startup Swanholm Technology is using TinyML to create workwear and gear for workers who are at a high risk for accidents and death. Its first product is a safety vest called the Alert Vest, which is designed to detect falls and send an alert when one is detected. The sensor sits on the base of a person’s neck while the battery back and radios are posited on either side. The electronics are packaged so they can be removed for washes, yet also rugged enough to withstand harsh environments such as outside in the rain and in mines. The case study published by Imagimob details the design considerations, training the model for fall detection and use cases. I loved reading it. (Imagimob)

Synaptics makes a TinyML acquisition for vision: Synaptics has acquired Emza Visual Sense, a company making ultra-low-power AI vision sensors. The acquisition provides algorithms combined with a low-power image sensor that will augment Synaptics’ existing portfolio of human-machine interaction products such as haptic sensors, touchscreens, voice, and more. Emza’s sensors will help provide context for electronic devices so they can detect presence and even gaze. This could help usher in new security applications on new devices, such as using your face to unlock your laptop or even a door. (Synaptics)

Node-RED startup gets funding: FlowForge, a startup founded to help commercialize the open source Node-RED programming tool, has raised $7.2 million to help enterprise and industrial companies use the framework to blend their disparate networks. Node-RED is a nifty tool that helps non technical users bridge different networks and create digital workflows. I have played with it in my smart home, but companies such as Hitachi and Siemens are using it in their industrial operations. There are other companies trying to make Node-RED easier for businesses to deploy and manage, but FlowForge has Nick O’Leary, one of the creators of the framework as founder. (TechCrunch)

Please stop trying to sell the industrial metaverse: First Nvidia and Siemens were touting it, and now, with a new report, Molex and Arrow are all about the industrial metaverse. Until we get easy-to-implement digital twins and factories with real-time connections to equipment and business and operations, we should cool it with the metaverse hype. Can’t we just focus on getting visibility into the manufacturing processes we have without having toit as something that’s not even moving the needle in consumer technology yet? (Molex)

Real talk from Bosch: I may not be the only person annoyed by all the hype around industrial IoT, 5G, and metaverses. Andreas Mueller, head of communication and network technology in the corporate research department of Robert Bosch, spoke at a 6G Summit and used the early hype around 5G to caution attendees about promising too much, too early with new technology. That being said, he added that he felt 5G for factories was now past the “trough of disillusionment” on Gartner’s famous hype cycle, and set to gain real adoption. I think next year is when we’ll see actual 5G end devices aimed at factory 5G deployments, with adoption following after that. Which will be about five years after the initial hype. Sigh. (Enterprise IoT Insights)

InfluxDB has re-architected its database storage engine: This is pretty nerdy, but time series database provider InfluxDB has rebuilt its core storage engine in Rust, replacing a previous version built in Go. The new storage engine was designed for the cloud and real-time access. This should speed up query access for data stored in the new engine and eliminates some of the restrictions on the time series data stored in the database. It also adds SQL as a programming language. Since time series data is essential for the IoT, having access to a faster and more scalable database is great. (InfluxDB)

This is a cool edge computing case study: Nvidia published a case study showing how it managed to replace the badges employees used to access its building with a hands-free option that uses biometrics and computer vision. The case study is really complete and shows how companies need to outline and engineer an edge AI project. For example, to get to desired latency times, Nvidia realized it needs to place its servers within 40 feet of the entrances. (Nvidia)