IoT news of the week for May 7, 2021

What are the real costs of using consumer tech everywhere? Consumer technology has democratized technology, taking high-cost components once used for military or rare use cases and turning them into parts for security cameras, for example. This trend also extends to manufacturers and industrial companies buying consumer tech to make their products smarter or cheaper. But given that most consumer tech companies envision their products lasting for three to five years, there’s a mismatch between how long a smart asset should live and how much it might cost to upgrade the electronic brains inside of it. I think about this a lot. For example, my seven-year-old Tesla just had its computer system replaced because the memory had worn out, causing it to lose vital functions, such as my speedometer and backup camera. This column argues that we should be thinking about the extended costs of this shift, and should either stockpile relevant parts or consider how to design systems for easy replacement. (IEEE Spectrum)

Will the blockchain help us with data integrity? Data integrity is a huge challenge for the IoT. If you build a system designed to take in sensor data, analyze it, make a decision, and then act on that decision, you need to know that the original sensor data or other input was legitimate. Put another way, you need to know that the data is from the device you think it’s from and that that device is performing as intended. One way to ensure that the data hasn’t been changed is by using the blockchain or another distributed ledger to make a record of the data at creation and track changes to that data if changes take place. I’ve discussed this idea before on the podcast, and I’m excited to hear how others think this would work. For it to apply to sensors we will need some sort of secure, low overhead method of writing data creation to a distributed ledger. Companies like IOTA are working on that. (Security Ledger)

Long-distance LoRa (like, really long-range) is coming: Well, satellite internet fans, long-range (LoRa) connectivity is coming to space. The popular Low-Power Wide-Area Network protocol has developed a new protocol called Long Range-Frequency Hopping Spread Spectrum (LR-FHSS), which will allow LoRa-based devices on the ground to communicate directly with satellites in the ISM band. Semtech, the company that makes LoRa radios, is working with EchoStar Mobile and the LoRa Alliance to create a viable LoRa network that can span space and terrestrial coverage. This would be a huge value-add for things like asset tracking in remote areas or on ocean-going vessels, and would be competitive with other satellite-based IoT networks that are only delivering direct satellite connections. (Semtech)

Timescale DB raises funding: Time series databases are the lifeblood of the internet of things. Or maybe they’re more like a reservoir of all of the data gathered by sensors. Either way, they are essential to IoT projects, so it’s no surprise that TimeScale DB, one of the providers of time series databases, has scored an additional $40 million funding round. What’s interesting is that the company has decided not to sell an enterprise version of its open source database, and will instead rely on hosting in the cloud for its revenue. (TechCrunch)

Smart cities enable digital authoritarianism: This article is worth reading because it lays out how prevalent surveillance tech has become in many cities, explains why that’s a bad thing, and details how the threat from smart cities extends even further than surveillance. It does all of this by pointing out the lack of security associated with many smart city deployments, which open them up to hackers and a loss of citizen data. It also touches on how difficult it can be to roll back such surveillance after the technology has been installed. I don’t think we will stop adding sensors to our urban areas, but I do think we should pause before adding more cameras and start a legislative process to protect civil rights associated with data gathered, stored, and analyzed by computers. (Foreign Policy)

Kneron raises funds and acquires image signal processing company: Kneron, a company building machine learning chips for edge devices, has raised an additional $7 million and said it would acquire image signal processing company Vactics. The Vactics acquisition is worth $10 million in cash; Vactics employees will work in Kneron’s surveillance and camera division. Kneron makes a chip that has a reconfigurable element, which can be optimized to run specific neural networks at low power. The six-year-old company has now raised more than $100 million in funding. (CNBC)

Trends in edge computer vision: Computer vision is becoming an integral element of smart factories, autonomous cars, and smart cities, so understanding the trends and new ideas governing computer vision is worthwhile. Take a moment to check out this article, which discusses the challenges and opportunities in embedded computer vision in the years ahead. (EETimes)

New optical tech for sleeker smart glasses: I won’t pretend to understand exactly what’s happening in this story, but the gist of it is that researchers at the University of Rochester have figured out how to build and make a thin layer of light-gathering film that will sit on a pair of glasses and help direct images to the eye. The tech is designed for delivering a compact AR or VR experience, so those of us who don’t want to wear a bulky computer on our face should follow the research with excitement. (University of Rochester)

A good review for the Philips Hue wall module: If you’re in the market for a light switch to govern your Philips Hue light bulbs, then you can opt for a wireless switch to replace your existing switch, or you can install a Philips Hue module behind your existing switch. The latter will be nicer-looking and match the switches throughout your home. However, you will have to do a little wiring. If you’re interested in that as an option, then check out this review (and some tips for wiring). (Tweakers)