Health-tracking dog collars claim to provide usable biometric data. Experts are preaching caution. – Mashable

What do smart home devices, activity trackers, and now biometric wearables have in common? They’re gadgets for your dog.
Following the rise of human wearables that track cardiovascular and respiratory health, several companies are suddenly promising the same in canine form. 
Three products all slated for release this year are all making similar claims, based on similar apparent breakthroughs. Using various forms of contactless sensors, these devices monitor heart rate (two also monitor respiratory rate), meaning they can theoretically infer a dog’s emotional state, or help detect heart conditions. Current smart devices for pets already come with built-in GPS and activity tracking, but the health monitoring aspect is new. Measuring a dog’s vitals with accuracy and ease has been a tricky problem. That’s why some experts are wary of the new claims. 
Adapting this technology for dogs comes with unique challenges. Sensors in devices like Fitbit or Apple Watch (ECG and PPG) require skin contact to get an accurate reading. Since most people probably aren’t about to shave their furry friends just to strap on a doggy Fitbit, these technologies have historically been a nonstarter.
That leaves contactless sensors like radar and acoustic technology, but those come with their own challenges. Translating radar signals into a coherent health metric is complicated. Acoustic signals require filtering out extraneous sound.
“We don’t have an electronics problem, we have a materials problem,” said Dr. Firat Guder who is principal investigator and chief engineer of his own research group in the Department of Bioengineering at Imperial College London. Simply put, it’s not the tech, it’s the application.
But several companies using contactless sensors say they’ve cracked it. French company Invoxia has developed a smart collar embedded with mini radar motion sensors that emit radio waves to detect subtle changes. An old technology that was just recently miniaturized, it’s the same technology used for movement recognition in Google’s Pixel 4. But instead of recognizing hand gestures, it’s tiny movements under the fur. 
“No matter the type of fur we will still be able to detect the movement of the skin and the speed of the movement,” said Amelie Caudron, CEO of Invoxia. “And from that, with some post processing, we’re able to extract the heart rate and the breathing rate.” Caudron says they filter out extraneous movements using their patented AI technology. Currently, Invoxia is conducting a clinical validation study with a third party that Caudron declined to disclose. 
In Taiwan at the Industrial Technology Research Institute (ITRI), Dr. Hong-Dun Lin is using doppler radar technology for their iPetWeaR collar. In a product test involving a small sample of 17 dogs and cats at the ​​​​Hsinchu City Animal Protection and Health Inspection Office, heart rate was measured by iPetWeaR and compared to ECG measurements. At 98 percent accuracy, the small, informal experiment showed promise. Lin, who has a PhD in electrical engineering said they are also testing iPetWeaR in collaboration with veterinarians at University of Taiwan and Chen Animal Hospital.
Instead of radar technology, Japanese company Langless uses acoustic sensors for a harness that claims to also detect your dog’s mood. The harness, called Inupathy, has a tiny microphone that records your dog’s heart rate and uses HRV analysis to supposedly measure your dog’s emotional state. HRV, which is also used in smartwatches like Apple Watch, is an indicator of how your body adapts to various environments. 
Like iPetWeaR, Joji Yamaguchi, founder and CTO of Inupathy said they compared its microphone system with an ECG monitor and achieved 90 percent accuracy. Yamaguchi, who has a background in system engineering, said they’ve developed Inupathy with veterinarians and dog trainers, and have tested it on hundreds of dogs. Yamaguchi also said that their data is being used for research at universities and laboratories in Japan. 
But until Guder sees the data, he says he remains unconvinced because he hasn’t seen any breakthroughs in the last two years that would indicate the existing challenges have been resolved. “The question is, what critical issues did these companies solve that could not be solved for a really long time?” He wrote in an email. “In other words, why now and not two years ago or five? Did whatever that led to their invention emerge recently?”
Guder, who is also an associate professor of bioengineering at Imperial College London, has been conducting research and development for sensor technology for humans and dogs for six years. It’s a niche area of study and he knows the landscape well. “As far as we know, these are really difficult problems. And if somebody solves it, it will be hard to convince people without showing the data.” 
In response to the skepticism, Caudron sounds confident about Invoxia’s product, and its proof of efficacy. “That’s quite interesting,” she said. “Because we do have results, and we do have this data on several types of furs and several types of dogs, etc. But let’s wait for the clinical validation, and then we can talk to them.”
Contactless motion detection is new, but Caudron says it is already being used in devices like Google Nest Hub (2nd Gen) to measure breathing. On the same note, Lin’s original idea was to use this technology for humans (fall detection, sleep evaluation, etc.) but pivoted when he realized its potential application for pets. 
Like Guder, veterinarian Dr. Ernie Ward looks forward to seeing the validation studies, but remains “scientifically skeptical” until then. “I’m still really skeptical until you actually see it, because it’s one thing to say ‘I can fly’ and another thing to actually see me fly.”
Ward, who includes evolving pet technology as one of his areas of focus, wants to see robust validation for such claims; that the product is “repeatable, accurate and demonstrable.” In a follow-up email, he wrote, “The tech is sound, but it’s the application and results that matter.” 
The stakes are high for any technology that promises health monitoring. What if a wearable wrongly detects a heart arrhythmia? What if it fails to detect one? Either scenario has disturbing consequences. Plus, FDA approval for animal devices is not required, and regulation for animals is generally lower. “If you’re going to invoke those types of claims, then we need some verification because the consumers just deserve it,” said Ward. 
Caudron says the health metrics gathered over time provide an objective baseline. So when it’s shared with your vet, they can make a more accurate diagnosis. So it’s not like pet parents will get a panic-inducing alert that their dog is about to have a heart attack. Caudron asserts that Invoxia is not meant to replace your dog’s vet, but it is still meant to be important. “The goal is to become like a cornerstone of the healthcare path for a dog, and to actually help the pet parents be more informed and go to the right service at the right time.” 
Other pet tech companies are taking a cautious approach to new advances in this realm. “We are expecting technologies measuring these vitals and others to emerge over the next few years,” wrote Jonathan Bensamoun, founder and CEO of Fi in an email. “And we will integrate them into the Fi Collar as soon as we consider them reliable,” he continued.
In an email statement, Wagz Founder and CEO Terry Anderton noted the obstacles in getting an accurate measurement through fur. While he declined to comment on future plans, he wrote that Wagz “has been working on other types of sensors that can reliably penetrate the fur to obtain accurate measurements, including the use of low frequency LiDAR (Light Detection and Ranging) and infrared sensor technology.”
Reliability aside, the question of why remains. How much of this data is actually useful to the average pet owner? “We tend to be chasing heart rate, which is super valuable, but I would much rather know weight changes, because we can really do something about that,” said Ward.
If we are not trained veterinary professionals, what are we supposed to do with this information? The answer seems to be, hand the information over to those who are trained veterinary professionals. They can decide whether or not it’s worth acting on, or worth anything at all.
Related Video: Move over, Boston Dynamics. You can build your own tiny robo-dog now.
More in Animals, Smartwatches

source