IoT news of the week for August 14, 2020

Honeywell is getting the collaborative spirit: Honeywell, which recently signed a partnership with SAP to help organizations understand the financial performance of their buildings, has signed onto another collaboration, this time with oil services giant Halliburton. The focus here is on both collaboration and bringing data from each company’s software to build new products and services. This is where the internet of things can really start to create value, because no one company will have all of the data necessary to document a business process. It’s worth paying attention to these deals, noting how well they actually progress and what customers adopt. That way we can track if customers are really buying into the promise of IoT and, in turn, delivering value as part of an ecosystem. (Seeking Alpha)

We’re finally ready to talk about standards for consumer wearables for medical-grade use cases: Silicon Valley has rushed to deliver wellness products in the form of activity trackers, sleep trackers, connected blood pressure cuffs, and EKG devices that run on your smartphone. But for the most part, these aren’t sufficient for medical-grade use cases because the data they produce isn’t standardized or consistently tested in peer-reviewed studies. With the rush toward the remote monitoring of patients and remotely conducted clinical trials, the medical device industry is calling for Silicon Valley to help corral the Wild West of Data that has emerged. In this essay, the authors call for companies to sign on to use standard, clinically validated methodologies for data collecting and sharing so researchers can compare results from patients wearing Fitbits to those wearing Apple Watches. Because while it’s fine if proprietary algorithms are used for the consumer-facing apps, for clinical trials, there needs to be more transparency. (Stat)

COVID cuts smart home revenue estimates: ABI Research says that in 2020, smart home device spending will rise by 4% from the prior year, to $85 billion. However, the new figure represents a $14.1 billion decline in predicted sales, since prior to the pandemic analysts expected spending to rise by 21%. This data actually surprised me because it suggests that COVID-19 has cut the adoption of smart home devices despite many of us spending far more time at home. Given that job losses and economic uncertainty make it hard to shell out $50 for a smart light bulb, and that expensive security cameras aren’t really necessary if you never leave home, I guess the data makes sense. But I also wonder how many people are, like me, tired of reading about privacy violations and devices that lose their functionality and/or have security flaws. (CE Pro)

It’s a cardiologist in a box? A startup called Bodyport has raised $11.2 million for its connected scale, which is designed to measure health. The scale is part of an overall medical-grade service that links a patient to their cardiologist and helps both track heart data without the need for in-person visits to a clinic. I’m assuming that because of the pandemic the company sees a chance to push its version of sensors and algorithms to the medical industry and the masses. But at this point I’m a little burned out by the mess of products competing to provide health monitoring without a focus on creating data standards and new systems for helping doctors manage what will be a huge influx of data. (Bodyport)

This is a fun-sounding startup: Las Vegas-based Terbine is trying to create a real-world market for data. This is a concept I’ve been excited about for more than a decade, but technical and logistical challenges seem to keep it forever on the sidelines. On the technical side, some of the challenges include trying to move large amounts of data and the speed at which some data becomes stale. For example, weather data may be relevant for a long time, but there’s a lot of it, so transferring it to another party comes with storage and associated bandwidth costs (the data gravity problem). Also, certain data — such as traffic light data — is only good for a few seconds or minutes, so a market would have to be faster than current ad exchanges. On the logistical side, most companies believe that their data is the new oil and want to hoard it, despite not knowing what to do with it or realizing that it would be more valuable when combined with data from other firms. So I’ll keep an eye on Terbine because, as I said, I love this idea. (IMC)

Apptricity releases a Bluetooth tracker that can go the distance: Apptricity has shown off a Bluetooth hub and tracker that can measure a location for 20 miles, far beyond the typical 30-to-100-foot range that Bluetooth radio waves can travel. The boost means that Apptricity Bluetooth devices can compete with SigFox, LoRa, and other Low-Power Wide-Area Networks for tracking devices across large corporate campuses and warehouse settings. (Apptricity)

Personalized medicine could be coming closer: One of the ideas celebrated by futurists is that AI and the IoT can help deliver medical interventions personalized for each individual. This requires pretty intrusive surveillance of your body, but if you could get a customized treatment for your cancer that killed the cancerous cells while minimizing any side effects, you’d probably take it. Now UCLA Samueli School of Engineering and the Stanford School of Medicine have built a smartwatch that analyzes a person’s sweat to detect trace amounts of medicine in their body. This would allow doctors to calculate the exact amount of a drug that would be effective for each patient — possibly reducing both the side effects and cost of treatment. (Medgadget)

This is a deep dive into new ways of performing machine learning on time-series data: For those of y’all who are familiar with convolutional neural networks, recurrent neural networks and other ways of training computers to extract meaning from data, this article talks about a new technique that combines different training models and is optimized for time-series data. They are called temporal convolutional networks or TCNs. This is relevant for IoT folks, because a lot of the data generated by sensors is time-series data, and more efficient ways to train algorithms to handle such data could benefit many IoT applications. This is a dense read, but worth it. (Medium)

SiFive just got a lot more money: SiFive, the company making it easier for designers to build RISC-V processors, has raised $61 million in Series E funding, bringing its total funding to about $190 million. The company is banking on a new open-source chip architecture called RISC-V that competes with MIPS, x86, and the Arm architecture. RISC-V is gaining ground because it doesn’t have the high licensing fees associated with some of the other options, and because designers can freely customize it for specific workloads. A few companies are using RISC-V chips to build chips for the IoT.  (HPC Wire)