I know that Philips Hue has/had a wall switch without a battery. This worked great.
No The technology is called SAW Sensors. The comparison for RFID was made as it is a more common technology that people have heard of as example for RF without batteries.
Idea is that you put a special device that resonates in the frequency on the window. so moving around/changing environment should be unproblematic
Reliability problems are there and might lead to a lost measurement. Which is not a problem if the device is stableā¦ ie. window is still open a second later / Button is still pressed 1/10th of a second later.
An Interplay with Videostreams may be less problematic than you think as I guess an ac WLAN is not taxed a lot by a videostream.
Replaying attacks are interesting, but hard to do, and we will keep an eye on that. But basically the main problem I see with these is to generate them correctly. Because the signal is analogous and not digital and power and frequency matterā¦ you might have to invest more than just a Raspberry pi to fake this well. I would have to talk to the engineers about how hard it actually is ā¦ but I doubt it can be done with a simple SDR. Similar with NFC payments, its hard to steal a few buck from someone using these. Jamming a WiFi / Zigbee network when you want to break into someones house seems to me the easier solution to bypass uncabled security.
It is rather the bleding edge of what is possible. But the people I work with have before cooperated for example with EnOcean to help make their system possible, so they are specialized in this sort of problem.
900 Mhz band is not a solution for this technology.
Iām not too sure about that. When you rely on measuring reflected interference patterns in an open space, which your device seems to be based on, things like multipath returns that create interference themselves can be very challenging, especially when theyāre moving and changing constantly. If you solved this problem entirely, well done. But Iād wait for an actual in-situ proof of concept myself 
A video stream (or any continuous tx stream) will literally drown out a channel completely. The only way for different APs to coexist on a same channel in such an environment is very clever error correction used by the physical transport layer in the Wifi standard. But such a system needs actual processing power on both sides of the link to manage the bidirectional communication in the crowded environment. I can see your technology work well in a relatively clean RF environment, but getting that to work in a center-city apartment with hundreds of overlapping wifi and people streaming stuff left and right would be very difficult, I imagine. Did you guys do any real life field testing in extremely noisy RF environments with several continuous high powered transmitters nearby ?
You keep bringing up these other technologies, but theyāre completely different from yours. NFC, like the RFID you mentioned earlier, uses active signal processing. An NFC tag contains a microprocessor that will process and actively send back data using a defined protocol. Itās just like a small battery powered device, only without a battery and powered by RF. Which requires either very short distances or very powerful transmitters. Your system is completely different. Itās like having several mirrors with different colored dyes that change depending on their state. You shine a white light onto them and see what color (frequency) it comes back to deduce the state of the mirror. Thatās all very nice and smart, but it has nothing to do with active RF technologies like RFID or NFC. Comparing to them is really counterproductive as far as I see it.
We like the idea of energy harvesting! It is kind of a specialty of ours. But usually the devices created are bigger and clunkier than other SmartHome devices. By Supplying the energy over the air, we wanted to have the benefits of the energy harvesting (no battery change) minus the clunkyness of the devices.
With Buttons energy Harvesting works rather well. Possibly the reliability is slightly decreased. The problem is that a door/ window contacts that I know of with energy harvesting seem still not as nice as the buttons.
You need a hub in each room + sensors. That is going to be very costly compared to what is currently on the market.
Multipath is not an issue as the delay in the answer of the device gets rid of the refelections from the sender.
We only tested in an office environment till now not with running videostream. I expect this to be problematic, but that a running Videostream for example from an UHD video will take about 15 MiB/s depending on your wifi this is less than a quarter of your Bandwidth. And most of this will happen in the 5 GHZ bands.
But yes this will need more testing.
It was my attempt to get the concept accross in a non technical manner. You are right of course that it may confuse people, but my hope was the opposite: to keep this non-technological and ask if someone was interested without confusing people and introducing the concept of SAW Sensorsā¦
Itās an interesting concept indeed. As for the price Iād be willing to pay, Iād say about 50 euro per room for 2/3 sensors + hub. Youāre competing with cheap Chinese sensors here. The battery-free aspect isnāt worth an extra 200 euros to me.
Ok, letās put all technical considerations aside and assume this technology works as advertised in my home. Letās only look at it from a practical point of view. What would the unique selling points of this technology be over existing wireless (battery backed) sensors ? Why would I want to buy this instead of a zwave/zigbee/whatever sensor, especially since itās pretty expensive.
As far as I see it, the only real advantage is that I would not have to change a battery every x amount of time. My zwave Aeotec door sensors need a battery change every 1 to 1.5 years on average. While annoying, itās not something I consider a major issue, especially since I get plenty of forewarning time and I typically keep a little stash of batteries around.
Is there something obvious I missed ?
Also I assume that you can only do sensors with mechanical actuation, like magnetic, temperature, humidity, etc. Motion sensors would not be possible. So I would still need another home automation system in parallel for those and for all active nodes like relays, dimmers, etc.
I can see the real value of your system in an industrial setting. But Iām not convinced that this is something that would bring enough value to the home automation sector at this point. Just my subjective point of view of course.