IoT news of the week for May 28, 2021

Is this an easier way to build digital twins? Digital twins are digital simulations of the physical world and a physical process. By applying real-world data to a virtual model people can understand how forces or stresses might affect an individual machine or building — or even see it writ large over the scale of an entire city. But building virtual models of the real world is time-consuming because building a model of a building is different than building a model of an engine. So I was intrigued by the news that researchers have figured out a way to mathematically model the state of a generic system, the data coming from sensors or observations within that system, and the options someone monitoring that system has to control or change it. Mathematically modeling these things provides a model that can be used to build a digital twin of anything. I don’t know how it works, but go check out the research, because if this is legit, it’s a big deal. (UT Austin)

IKEA’s plans for smart home privacy include AR: One of the biggest challenges in data privacy is letting people know what data is collected and how. IKEA’s R&D lab, Space10, is showing off a way to provide such transparency with augmented reality. Using AR overlaid on common smart home devices, developers could show how a TV talks to a doorbell or how your Google hub talks to your locks, for example. IKEA’s designers also looked at using computational audio to create quiet spaces in the home and using a blockchain to store data about smart home devices’ provenance, the carbon footprint, and their security practices. This is an idea already used by the Matter standard (formerly Project CHIP) for security, but it could easily be expanded. (Fast Company)

More details on wireless sensing technology: I’m super excited by radar, ultra-wideband, and Wi-Fi for sensing the placement of objects and people. Broadly speaking, those technologies measure disturbances in emitted radio waves to figure out what’s in front of them. Each tech works differently, but all of them are becoming more important for the smart home and smart factories. That’s why I was stoked to read about Celeno’s new chip, which combines Bluetooth, Wi-Fi, and what Celeno calls “Wi-Fi Doppler Radar.” Celeno describes the Wi-Fi Doppler Radar as a combination of the Doppler effect and standard Wi-Fi packets. Using the tech, an algorithm can track an object’s location or sense its presence or motion. The technology requires only a single Wi-Fi device as opposed to multiple devices around a room. I’m currently working on an article about these sorts of technologies, so stay tuned for more. (New Electronics)

Edge Impulse gets $15M for tiny ML: I can’t believe I missed this last week — especially given my love for both tiny ML and the work that Edge Impulse is doing to make it easier to build tiny ML models — but the startup raised $15 million in Series A funds led by Canaan Partners, with Acrew Capital, Fika Ventures, Momenta Ventures, and Knollwood Investment Advisory participating. I’m glad everyone else is getting excited about running machine learning directly on embedded devices and less powerful processors, because it’s going to change the way we experience the world around us. It will provide more privacy, faster response times, and could mean that not everything needs a radio. (Edge Impulse)

Intel’s report on edge computing is worth a read: Yes, this report is all about how much processing power will be needed at the edge, which should theoretically be a bonanza for Intel’s chip business, but its thesis rings true. Plus, it’s chock-full of interesting startups and tidbits from case studies about how companies are using edge computing in their business. Check it out. (Intel)

Twilio’s eSIM Super SIM is good for IoT: Twilio is making it easier for companies to put more flexible eSIMs into their devices. eSIMs are software-based SIM cards that provide connectivity without requiring a physical SIM swap to change carriers. With Twilio’s Super SIM service, the makers of devices can choose what carrier to switch to when using an eSIM and manage the process themselves. Historically, customers relied on roaming to get connectivity across different carriers, and roaming was at the mercy of the carriers themselves. This update gives device makers more control over their connectivity and does so in a manner that lets them avoid having to swap out physical SIM cards. It’s a win-win. (Twilio)

Get ready for injectable chips: I probably shouldn’t feed the conspiracy theorists, but researchers at Columbia University have managed to develop a chip as small as a dust mite that’s capable of being injected into a patient. The chip measures body temperature and uses ultrasound waves as a power source, as well as to communicate with an external device. The choice of ultrasound was determined by the size of the chip. At 0.1 mm³ it was too small to effectively conduct larger radio waves. (Medgadget)

How to determine if the EU will trust your AI: The EU is taking up the mantle of privacy and AI regulation with a vengeance even as federal regulators in the U.S. keep punting on the topic, forcing states to take control. This month, the EU published the results of an inquiry begun two years ago on the topic of trustworthy artificial intelligence, including seven steps to determine a trustworthy AI. The ethics guidelines are pretty standard, calling for designers of AI algorithms to have systems in place for measuring how they perform and that they consider diversity and environmental goals, that the use of data and collection of data be transparent, and that humans and backup systems need to be in place in case of problems. Click through to read about these in more depth. (EU)

ARM expects the Nvidia deal to close in April 2022: And while it’s waiting for the deal to close, it has implemented a hiring freeze. (The Register)