For a brief, beautiful moment from the headline I thought "oh good, more PoE sensors finally happening!" doho. But no not just about battery but any power at all, and an interesting approach for further research. Of course to actually get data from still requires powered devices but in a lot of cases it'd be much easier to have that be a single or small number of more easily placed central units vs every single sensor tag separately. Ultra low cost and simplicity are values all their own in terms of applications.
My biggest immediately question though and one I'm a little surprised not see addressed, even at the research stage, is any mention of other animals. There is a bunch there about the ultrasonic frequency being well above the human limit of ~20 kHz. But IIRC for example dogs can hear up to like 45-60 kHz, and cats all the way to 65-85 kHz. I assume lots and lots of other animals also can perceive sounds well beyond human senses. Noise pollution is already a somewhat unrecognized but big problem for all sorts of life around us (not that it's irrelevant to human health either), so if more use of ultrasonics made that worse that's a concern. And as a practical matter the product market is probably going to shrink pretty dramatically if it drives pets mad, a lot of people have pets nowadays they care about a great deal. For that matter even in public environments if it messed with service dogs that might have ADA or equivalent implications.
Still, good reminder of various side channels one doesn't always think about.