IoT news of the week for March 4, 2022 – Stacey on IoT

The Victor Smart-Kill Electronic Mousetrap. Image courtesy of Kevin Tofel.

Kevin tries out a Wi-Fi mousetrap: Got rodents? Kevin does, so he purchased a Victor Smart-Kill Electronic Mousetrap for $39.95. The trap doesn’t talk to Alexa, but it does send a notification after it zaps the mouse so you don’t have to keep checking an empty trap or leave any dead animals to rot. See what he thinks. (Stacey on IoT)

SmartThings is going after the multifamily business: Samsung SmartThings has teamed up with CommScope Ruckus to put its home automation software inside Ruckus’ Wi-Fi equipment aimed at the apartment market. SmartThings and CommScope Ruckus also worked with Picerne Real Estate Group, a construction and development firm, to create connected experiences in one of Picerne’s new apartment developments. Adding smarts to apartments helps landlords by letting them eliminate keys throughout the property and control thermostats while an apartment is empty. Residents also tend to want smarter apartments, although I feel like more and more people move with their own wireless access points rather than rely on one provided by the landlord. (SmartThings)

OpenSpace gets $102M for smart construction: Adding cameras and other sensors to construction sites has been a big trend that can help speed up the building process as well as cut down on waste and injuries. OpenSpace relies on cameras to snap overview pictures of a construction site and then analyzes those pictures for information about how the process is going. It’s one of several startups applying AI to construction, and it has raised $102 million from a group of investors including Penny Pritzker’s PSP Growth fund, Black Rock Capital, Menlo Ventures, and a unit of commercial real estate brokerage Jones Lang LaSalle. (Bloomberg)

This video doorbell is also a lock: Eufy has launched a Kickstarter for a $399 video doorbell that doubles as a smart lock. This isn’t the first combo device of this type that I’ve seen, and I’d look at it only because I’d like one less device to manage in my home. The lock opens with a fingerprint scan or a keypad. This might be odd for visitors, because the same place you press to ring the doorbell is also the place you’d press to unlock the door, and in this case the doorbell is where your old lock used to be. There’s also the question of battery life. This is a Wi-Fi video doorbell and lock, and moving the deadbolt and recording motion is going to take a toll on the batteries. The lock/doorbell will ship in May, so maybe we’ll test it out. (The Verge)

Cool research to prevent side-channel attacks on IoT devices: MIT researchers have figured out a way to secure locally processed data on a chip to prevent hackers from messing with the data processing that happens on said chip. This sort of research will become increasingly important as more decisions get made locally on the hardware as opposed in the cloud. Figuring out effective ways to secure power-efficient and constrained silicone is going to be a very big deal, so it’s good to see more options. (SciTech Daily)

Trusting AI models in medicine may require tighter controls: This is an eye-opening study from STAT News and MIT that tracked how AI models performed over time when it came to figuring out certain types of medical risk factors. The algorithms started out strong, but degraded as time went on. The interesting bit is why they degraded. It appears that they became less effective because some of the medical codes used in the software platform changed and because the way certain samples were taken also changed. The first issue means data scientists and medical software providers will have to consult on updates and reweigh their models as the software changes. The second is an indication that human performance and data gathering has the potential to influence an algorithm, and as the human performing a job changes over time, the results can also change. The problems causing the AI ​​models to degrade are intuitive but really hard to control for, and mean that our super scalable AI future still has some clear limits and roadblocks to overcome. (STAT News)

Vodafone launches a smart contracting platform for connected devices: Vodafone has launched a digital asset broker it calls the Economy of Things platform that will let devices trade with each other. The idea is that if we want to let connected and AI-enabled devices interact with each other, they need a platform to help establish what devices can be trusted and what assets a device might have backing it. So if your autonomous vehicle wants to charge at a connected charging station, there’s a trusted platform to link the car’s financial information to the charger trying to get some value in exchange for the electricity. This speeds up transactions between devices and eliminates the wait for manual payment. (IoT Tech News)

This is a cool video showing how we may communicate with ambient intelligence: Google’s Advanced Technology and Projects (ATAP) group has created a video that shows how we may one day interact with homes full of intelligent devices. I’ve written that we’ll need new forms of interaction as we work towards ambient intelligence in the home. In the video, Google shows how it created a framework with a set of four movements, called Approach, Glance, Turn and Pass. Each action can be a trigger for behavior from a smart device such as a connected thermostat or a display. We’ve also seen Google use its Soli radar to implement some of these movements in existing devices, and if we see Soli embedded in more places, they could become more familiar for us. Check it out. (Engadget)

Some additional data on the number of devices we are adding to our homes in the US.: I like to keep an eye on the number of gadgets connected to home Wi-Fi networks because it helps me track the increasing number of smart televisions, tablets, appliances, doorbells, etc that are gaining ground in consumer homes. Plume, a company that provides Wi-Fi management as a cloud service for ISPs and for Wi-Fi vendors, has tracked the rising number of connected devices in homes in the US and Europe. In December 2015, homes had an average of eight connected devices, which rose to 19 by December 2021. Most of those devices are phones, laptops, printers and other well known products, but the average number of connected IoT devices has increased from 4.2 in December 2015 to 12.5 in December 2021. Meanwhile, I think I have about 60-80 Wi-Fi devices on my network. (Plume)

Leave a Reply

%d bloggers like this: