Ultrasound can cause anomalies in the Xiaomi MI5S Plus gyroscope(medium.com)
medium.com
Ultrasound can cause anomalies in the Xiaomi MI5S Plus gyroscope
https://medium.com/@juliodellaflora/ultrassom-pode-causar-anomalias-no-girosc%C3%B3pio-do-xiaomi-mi5s-plus-4050d718bc7f
44 comments
This has the potential to be dangerous if it can interfere with the raft of self-balancing 'personal transporters'.
This is exactly why we need to escape the mentality of sensor fusion and make our robots reliable just based on optical data. Then they would be susceptible to attacks that would affect humans too.
Humans use multiple sensors together at one. Your inner ear is a great example. If you see a picture doing one thing, but your inner ear feels another, this is the senstation that causes motion sickness. I cannot find the source right now, but i recall being taught that research dating back to US airforce simulation training suggested that this might be a built in evolutionary function that is designed to make us throw up. The basic narrative being "I'm seeing something I'm not feeling, i must have eaten something that's poisoned me, i should throw up so i don't die".
Edit: As a side note, this is why I'm extremely skeptical of things like VR technology in general (as a person that owns a HTC Vive). I'm not sure we're going to be able to easily convince the human brain it's OK for different sensors to feel different things, and not result in us feeling sick.
Edit: As a side note, this is why I'm extremely skeptical of things like VR technology in general (as a person that owns a HTC Vive). I'm not sure we're going to be able to easily convince the human brain it's OK for different sensors to feel different things, and not result in us feeling sick.
Re: your side note
That's why even though I'm a relatively avid gamer with plenty of disposable income, I've never bought on to the VR stuff. It's not about the cost, or the rapidly improving technology, it's that I don't believe (long term) that I'm going to really enjoy many VR games.
I have a racing wheel because I like to play racing games, but most of the time I end up using my controller anyway. Why? Because it doesn't feel real. The wheel ends up feeling farcical because yeah I'm cornering at 1.5G, but I don't feel a thing. In real life you're less likely to corner too fast because you can feel that you're going too fast. You're being shoved around, you're being moved side to side, you're sliding forward as you brake and backwards as you accelerate. In a video game you get none of that, and you become detached.
I think there will be plenty of valid VR experiences, once we figure out what's the right fit. Windows Mobile took the desktop and put it on a phone, it wasn't until Android and iOS came out that we figured that smartphones are really actually quite cool, provided you build the right interface. We now know there are just things you can't (or shouldn't) do on a mobile device compared to a desktop, and we build around that.
I don't know if playing an FPS in VR is ever going to feel "right" because you're entering the uncanny valley of "I'm walking up the stairs but I'm not actually". At least on Counter Strike you're just pushing buttons on a keyboard and looking at a small rectangle, so your brain knows it's not actually supposed to be happening.
That's why even though I'm a relatively avid gamer with plenty of disposable income, I've never bought on to the VR stuff. It's not about the cost, or the rapidly improving technology, it's that I don't believe (long term) that I'm going to really enjoy many VR games.
I have a racing wheel because I like to play racing games, but most of the time I end up using my controller anyway. Why? Because it doesn't feel real. The wheel ends up feeling farcical because yeah I'm cornering at 1.5G, but I don't feel a thing. In real life you're less likely to corner too fast because you can feel that you're going too fast. You're being shoved around, you're being moved side to side, you're sliding forward as you brake and backwards as you accelerate. In a video game you get none of that, and you become detached.
I think there will be plenty of valid VR experiences, once we figure out what's the right fit. Windows Mobile took the desktop and put it on a phone, it wasn't until Android and iOS came out that we figured that smartphones are really actually quite cool, provided you build the right interface. We now know there are just things you can't (or shouldn't) do on a mobile device compared to a desktop, and we build around that.
I don't know if playing an FPS in VR is ever going to feel "right" because you're entering the uncanny valley of "I'm walking up the stairs but I'm not actually". At least on Counter Strike you're just pushing buttons on a keyboard and looking at a small rectangle, so your brain knows it's not actually supposed to be happening.
I think you and I are in a similar thought pattern here. I only see a path forward for VR if those crazy treadmill / fan type setups become super mainstream and perfectly reliable. Even then, falling will never work! It's kind of crazy we want to simulate reality so badly when we're already in it.
Humans fuse a lot of sensory data though. For example, walking without proprioception is incredibly hard: https://philosophyandpsychology.wordpress.com/2007/10/14/liv...
I have nerve damage because of a spinal problem, I walk with a slight limp not because I've lost strength but because I can't always accurately 'feel' where my foot is.
It's like walking with pins and needles.
It's like walking with pins and needles.
I took this as an argument on why we shouldn't rely on one type of sensor alone, as it's easier to fool with tricks like this.
Humans with issues in their vestibular system located in the ear can't really use their optical data streams to compensate. When you loose your gyros time to get a walking stick or a beautiful nurse to lean on.
Interestingly, it is possible to learn to walk when you never had a vestibular system, although the gait is visibly distorted and when you know what you are looking for you can easily spot it. Those who suffer from CHARGE, and are left able to walk, can still do so despite sometimes entirely missing their vestibular system. But learning to walk with a vestibular system and then losing it later seems to be very, very difficult to recover from, if it is possible at all.
Can humans with issues in the optical stream use the gyros? I find it harder to balance with my eyes closed, though obviously not impossible.
Well, the blind and sight-impaired have been walking (and running) for years, so I'd vote "yes".
Even people with impressive visual hallucinations are still capable of walking.
That said, vertigo (where your eyes and your inner ear are in violent disagreement as to whether you are moving) can impair movement, though you can grow accustomed to the "disagreement".
Even people with impressive visual hallucinations are still capable of walking.
That said, vertigo (where your eyes and your inner ear are in violent disagreement as to whether you are moving) can impair movement, though you can grow accustomed to the "disagreement".
I think the optical stream fits more of an error correction/secondary data stream in this case.
The difficulty may be down to you trying to use other senses and overcompensating on the data (e.g. air pressure, which tells you something is near-ish you but not if it's about to contact you so you react)
The difficulty may be down to you trying to use other senses and overcompensating on the data (e.g. air pressure, which tells you something is near-ish you but not if it's about to contact you so you react)
To be fair, you trick an optical sensor with a gust of wind and a plastic bag.
Ideally all sensors would be redundant, with poorer and poorer (but still safe) performance as each is removed.
Ideally all sensors would be redundant, with poorer and poorer (but still safe) performance as each is removed.
Afaik equilibrium is a matter of at least vision and audition, is it not?
Topic from a year ago - "WALNUT: Acoustic Attacks on MEMS Sensors": https://spqr.eecs.umich.edu/walnut/
So if you don't want your neighbour flying his drone over your property can you shoot it with ultrasound? What would the legalities of that I wonder?
over your property
How much airspace do you own over your own property?
How much airspace do you own over your own property?
IIRC it's all the airspace until the 'general aviation' zone starts, where the owner still has some control but FAA or other local authority legally requires property owners to permit operation of planes and other properly licenced aircraft.
There's not much court precedent yet, and many places are passing specific laws to clarify this, but the general assumption should be that unless you're above 400-500 feet where national aviation regulations apply, you'd need a permission from the owner of the land.
There's not much court precedent yet, and many places are passing specific laws to clarify this, but the general assumption should be that unless you're above 400-500 feet where national aviation regulations apply, you'd need a permission from the owner of the land.
Could I destroy an RC vehicle driven onto my land property? The laws should be the same for both, although I'm not banking on it.
The general principle seems to be that things owned by others remain so even if they are on your property, and intentionally damaging them is vandalism. E.g. if you notice someone's car parked on your land and smash its windows, that'd be a crime even if that car "has no right to be there" - you could require the car to leave in some legal manner, possibly have it towed or sue for damages caused by it being there; but breaking other people's stuff is one of the oldest things prohibited by law.
Could you? Sure! Could you legally destroy it? As they say, "possession is nine-tenths of the law."
I read "over your property" as already implying that the airspace isn't your property.
I honestly didn't/don't know. How far do you own down?
Depend of the country, in France it's to the center of the earth, which means you receive some money when they build a metro under your house becaude they buy a right to use part of your property.
Iceland invaded by French Marines; geothermal energy in magma was "not theirs to take", says Macron"
In UK I'm pretty sure it's nothing, certainly you don't get anything (money/rights) if oil or coal is extracted from beneath your property ... but then people are allowed - with planning permission - to build down from established property. I expect the Crown owns it all.
Destroying other people’s property is usually illegal. I don’t think it becomes allowed just because it’s somewhere you own.
But the ultrasound isn't damaging it. It's just causing interference. That's like saying my radio is damaged when it gets interference from my mobile.
It's not actually damaged, it could be damaged from the fall but you didn't create gravity.
It's not actually damaged, it could be damaged from the fall but you didn't create gravity.
Hmm,I wonder ...
In UK I gather you're allowed anti-climb paint - which damages clothes - to prevent unauthorised access (trespass, say); but you're not allowed glass or spikes which damage people too. People usually signpost use of deterrents, I'm not sure that's a requirement, probably just part of the deterrent.
So there's certainly precedent for damaging others property to hinder "trespass".
You're generally allowed ultra-sonic devices like cat scarers that deter people's pets from gaining access to your property, and whose effects impinge across property boundaries.
So I reckon as long as you signpost "ultra-sonic deterrent" then the onus is on those who're infringing your privacy and peaceful enjoyment of your property to avoid any potential damage by avoiding your property.
Aside: Not sure how far the law has come but there are guidelines from CAA saying not within 50m of another person, not within 150m of built-up areas (eg https://publications.parliament.uk/pa/ld201415/ldselect/ldeu...).
In UK I gather you're allowed anti-climb paint - which damages clothes - to prevent unauthorised access (trespass, say); but you're not allowed glass or spikes which damage people too. People usually signpost use of deterrents, I'm not sure that's a requirement, probably just part of the deterrent.
So there's certainly precedent for damaging others property to hinder "trespass".
You're generally allowed ultra-sonic devices like cat scarers that deter people's pets from gaining access to your property, and whose effects impinge across property boundaries.
So I reckon as long as you signpost "ultra-sonic deterrent" then the onus is on those who're infringing your privacy and peaceful enjoyment of your property to avoid any potential damage by avoiding your property.
Aside: Not sure how far the law has come but there are guidelines from CAA saying not within 50m of another person, not within 150m of built-up areas (eg https://publications.parliament.uk/pa/ld201415/ldselect/ldeu...).
Same as shooting it with a gun, modulo local ordinance.
Is that really the same? Can you fire off a gun in your backyard at the sky? Isn't that illegal as the bullet comes back down to earth it can kill.
Meanwhile ultrasound is just lot's of sound waves and doesn't go far, and can't kill anything. I can fire ultrasound up into the air without a license or it being illegal.
Meanwhile ultrasound is just lot's of sound waves and doesn't go far, and can't kill anything. I can fire ultrasound up into the air without a license or it being illegal.
But doing it intentionally, whether as a triggered booby trap or an aimed shot, and taking down a drone would likely still be illegal if shooting a drone down with a gun is illegal.
I can’t remember for sure whether it is illegal to do so over your property, but I think it is.
I can’t remember for sure whether it is illegal to do so over your property, but I think it is.
Are you allowed a high fence (assuming neighbours don't object)? What if it's virtually invisible, like nylon fishing line? What if the fence is incorporeal, like an ultrasound "barrier".
Fence height restrictions are quite common. Not sure about whether they are limited to "visible" fences. I doubt it.
Use a defensive drone that fires vertically upward into the airspace over your property?
> can't kill anything
Unlike a large drone fully powered up hitting an unsuspecting person.
Unlike a large drone fully powered up hitting an unsuspecting person.
Or you can use a stick
https://www.youtube.com/watch?v=uhbuEUoxntQ
https://www.youtube.com/watch?v=uhbuEUoxntQ
Depends, not all drones use gyros.
All drones do use gyros - they couldn't remain level without having rotational awareness. Accelerometers are the sensors which are optional.
The earliest drones used actual model helicopter gyros, the current generation have those sensors soldered onto the control boards.
The earliest drones used actual model helicopter gyros, the current generation have those sensors soldered onto the control boards.
Not all drones remain level. Racing drones rely on operator input to remain level (it's called acro mode).
The PID cycle that controls the rotors, even on racing drones, relies on the gyros to provide the feedback potion of the PID loop. That feedback is "what is the current yaw, pitch, and roll of the drone".
Even in acro mode (which is more accurately defined as "the sticks determine the yaw, pitch and roll rotational velocities"), the gyros are used to measure the actual rate of rotation so it can be matched against the requested rotational velocity.
Positional stabilization, something you see more in the higher end drones, have an additional PID loop that takes input from the gyros, accelerometers and even GPS to determine the input to the underlying control PID loop.
Even in acro mode (which is more accurately defined as "the sticks determine the yaw, pitch and roll rotational velocities"), the gyros are used to measure the actual rate of rotation so it can be matched against the requested rotational velocity.
Positional stabilization, something you see more in the higher end drones, have an additional PID loop that takes input from the gyros, accelerometers and even GPS to determine the input to the underlying control PID loop.
And (roughly) Acro mode needs a gyro to control that the rotation matches the user request, whereas self-leveling needs an accelerometer too to figure out where "down" is.
Whtas the need for an Ultrasound. Are we now going to test our own body with such a system?