Im not sure I got it. It generates acoustic signals and a microphone server picks them up.

If so, will it penetrate through walls?

Trivia time! The first practical wireless remote control for TVs also used ultrasound: https://www.youtube.com/watch?v=MLPk1Us62xQ

It's a unique idea that I could see being useful in select situations. The reliance on wearable microphones sounds like a downside.

Also I guess this might be annoying for pets that can hear well beyond 20 kHz.

I can see some uses, but calling this system batteries free seems a stretch. A sensor is worth nothing if it can't be read, and to read this you need a powered microphone and computing. Some already common magnetic door systems do the same; door plate and magnet movement is enough to create a detectable current, (using no external power), then that signal is read and computed by an electronic/digital system (using power).

There already exists "kinetic" switches for lights etc whose switch contains some passive electronics that when actuated produces enough energy to emit a radio signal that can be read by a relay module. They're pretty handy as you can basically place the switch anywhere you want without the need for the wires to be there. The relay can live in the light fitting or somewhere else convenient.

There's probably no reason why these kinetic switches can't also be used for detecting other events like doors opening/closing etc. I feel like a radio signal is a bit more reliable and easier to detect than high frequency sound.

I also think calling these a "sensor" is a bit of a stretch. They detect events but have no knowledge of the current state of the thing they're sensing. E.g. the can detect a door opening/closing, but have no idea if the door is open or closed at a given time

Would these maintain their signature under repeated use? Or regular wear and tear?

It's research, not a product. Even with that, framing it as a smart home sensor in the press release is a stretch.

1) 93.75% success rate in controlled conditions, 92.1% in a somewhat-realistic deployment scenario - too low for reliability. I wouldn't use something like that to trigger smart home automations.

2) Range hardcapped at ~1m due to how ultrasound works, you can't centralize detection. Their answer is to give everyone in the household a wearable receiver, which is eeeeeeeh idk, doesn't look consumer-friendly to me.

3) Paper suggests a mix of durable and consumable parts for the transmitter. Their numbers show that the 3d-printed PLA cantilever needs to be replaced every 900 cycles or so. Should work fine, but...

4) ...every transmitter pair needs to be tuned per-setup, every time. Not a plug&play in the consumer sense.