You are here

A drone with a tele-doctor and defibrillator in 100 seconds?

There's a lot of excitement about the potential of autonomous drones, be they nimble quadcopters or longer-range fixed wing or hybrid aircraft. A group of students from Singularity University, for example, has a project called MatterNet working to provide transportation infrastructure for light cargo in regions of Africa where roads wash out for half the year.

Closer to home, these drones are not yet legal for commercial use, while government agencies are using them secretly.

Here's one useful idea: A small set of medical drones scattered around the city. Upon emergency call, they can fly, via a combination of autonomous navigation and remote-human-operated flying at the end, to any destination in the city within a couple of minutes. Call 911 and as soon as you say it's a medical emergency the drone is on the way. When it gets there, the human operator lands it or even sends it in a balcony on tall buildings with balconies. Somebody has to carry it to the patient if they are far from the outside.

When it gets to the patient it has a camera and conferencing ability to a remote doctor can examine the patient and talk to people around the patient to ask them questions or give them instructions. It also could contain one of those "foolproof defibrillator" modules able to deal with many kinds of heart attacks. They are already in many buildings but this way they could be anywhere. It's more useful than a taco.

The remote doctor could advise any medical staff who come, or give advice to the ambulance that's on the way but not getting there for a few minutes. If a medicine that can be administered by a layperson is needed, there might be some in the drone but a second drone could be loaded and dispatched within a few minutes as well -- that might take longer to fly but less time than an ambulance. You might not put any valuable medicines in the first drone to prevent people from summoning them just to steal them, though this might just happen for the valuable drone unless steps are taken to make that non-productive.

This should be combined with something I have felt is long overdue in the world of our mobile phones. People who are able to be on-call EMTs and doctors should have their phones updating their locations with a medical service while they are on call for such action. Then anybody with an emergency should be able to summon or get to the closest professional very quickly. (Of course there is no need to record this data after it changes, to avoid making a life-log of the doctor.) Nobody should ever have to ask "is there a doctor in the house?" 911 should be able to say, "There is a doctor 3 doors down, she's been notified." But the drone can always come, and bring a remote specialist if need be.

The other barrier to this is network dead zones. A map would need to be made of network dead zones and the drone would not fly into them, though it could fly through them. It would land just outside the dead zone and warn people not to carry it into one if the remote doctor's services are needed.

Someday, the drone could contain a winner of the X-prize "Medical Tricorder" contest with sensors to diagnose all sorts of conditions, and it might even eventually be a robot able to administer emergency drugs -- but the actual delivery and video feed is something we can do today.


Love it! Great thought. It would be difficult to set up but completely worth it. And you're right, so much better than Tacos.

Part of what limits AED availability is that their cost is somewhere in the ballpark of $800-$1200 for a low end model, and the probability that having one in the average home would save a life makes putting one in every home absurdly expensive in terms of price per life saved, if nobody in the home has risk factors that would suggest an AED would help.

If you want to solve the AED availability problem, figuring out how to build one for $20 and make it at least 80% as reliable as one that is fully FDA approved might be just as effective as an AED delivery drone. I suspect a good chunk of the $800-$1200 is the government trying to guarantee that people will die because the AED was too expensive instead of dying because it was unreliable (when it might not have been unreliable anyway), and we're probably killing people by worrying about the wrong things.

I also think a smartphone accessory might be a cheaper approach than a standalone AED. Does the AED do anything other than read heart rythm (borrow the audio in on the headset jack, or maybe just set the phone down on the heart and let its internal microphone do the work), charge a capacitor, and then deliver a shock at the right time? (For that matter, could Googarola decide to just build an AED into every smartphone if they could figure out how to do it for $5? And could you use the smartphone's radio to allow a doctor to run it in the manual modes remotely as well?)

Joel, you're right in that a significant part of the approx. $1K price of an AED is to make sure it is as close to 100% reliable as possible, when something that is 95% reliable and 1/2 the price might actually end up saving more lives. But it's ethically easier to provide fewer life-saving devices than many devices that might fail every now and then end up not resuscitating someone.

This is especially true in today's litigious and media-blasting society. Something goes wrong on one patient, and you're required to perform a costly recall on all devices sold, not to mention the PR damage. The FDA just responds to that, they recently put the clamps down on all defibrillators because some implantable defibs were found defective. Trying to recall an implanted device is not an easy matter!!! So they won't allow a cheap, 95% reliable device on the market, and you can't sell any medical device in the U.S. without FDA approval.

Clearing up some defib technicalities - the heart rhythm analysis is done by detecting the electrical rhythm of the heart, not the auditory sound of the blood. And cell phones don't have anywhere near the battery power to charge the caps up enough for a proper shock (usually 200 joules for a full jolt). The smarts could be in a cell phone, but the battery and capacitor and electrodes would have to be an additional unit.

Regarding the drones, this is a neat idea. But I think the medicine would be quite limited, since most would require a trained practitioner (anything with a hypodermic needle) or take enough time to deliver the effect that waiting a couple minutes for the EMT wouldn't matter. Mostly the advantage would be to get advance video to a doctor or nurse who could then prep the EMTs so they prepare and carry the right equipment. Oxygen sensor, ECG data, that would be fairly easy to provide as well.

But these drones would need to be able to fly in a city environment without going high enough to interfere with air traffic. Something capable of landing on a balcony implies a helicopter design, but a drone helicopter might not fly fast enough to beat the EMT team, unless there was one stationed every couple blocks. This might be more useful for backcountry rescues for skiers and hikers (and a fixed-wing drone, just basically do a controlled crash land).

It seems that 200 joules means that if you're willing to spend 40 seconds charging a capacitor, you only need to be able to provide 5 watts for those 40 seconds.

A typical cellphone battery seems to be 850 mah, which apparently works out to about 11000 joules, or 55 of those shocks if the battery happened to be fully charged at the time of the heart attack.

If draining a cellphone battery at 5 watts for 40 seconds turns out to be abusive in terms of shortening the battery life, that's OK, as long as it doesn't create an immediate fire hazard. Replacing a $50 battery after AED use is a totally acceptable expense (it's a lot better than spending $1k for an AED that is never used). And if the current battery chemistry doesn't support this sort of usage, but there's a way to spend an extra $1 per battery pack that would enable this usage, building better batteries may make sense.

Even if current AEDs detect electrical rhythm, is it actually necessary to rely upon electrical rhythm instead of audio? Presumably AEDs were originally designed with the idea that they would have special purpose hardware, and an electrical pickup is no more complicated than a microphone, given that a standalone AED has no other use for a microphone. Smart phones have a substantial amount of DSP capability, and there are even accelerometers which might help if you need to pick up the sound at different points and know something about how the phone was being moved as the sounds change, and if you can simplify the dedicated hardware, the cost drops, even if the software requires some significant effort to develop.

If getting FDA approval is hard, could this be deployed first in another country? The FDA would be in an interesting position if a news reporter could tell the story of someone who had some particular model of smartphone who happened to be near a person who might have benefited from an AED where no approved AED is available, if a representative from the smartphone maker started talking about how that model included a cheap AED in some other country, and has saved lives in that other country, but wasn't included in the US version due to FDA safety concerns. (I'm pretty sure the introduction of 911 and civilian access to GPS both followed tragic stories...)

I work on defibrillators, but I'm not an EE so I'll do the best I can. We need high voltage capacitors to deliver the waveform needed to shock the heart. This is a fast, biphasic pulse lasting about 10-20 mSec, so it needs to be high voltage. The battery can't work this fast, so the energy needs to be stored in a capacitor. Supercaps are good but they don't handle high voltages, so we're left with normal capacitors, which are pretty big. (We're talking 6" long, 1" diameter at the very least). You wouldn't just build that into everybody's smart phone. Then the stick-on pads to put on the chest and a couple feet of wire will take up more room.

You're right about the 40 second charge time, and that's not terribly bad. Our defibs normally charge in about 8 seconds, but in low battery conditions could take over 20 seconds. (It does the charging while analyzing the rhythm, which needs at least 6 seconds of ECG data). I'm surprised a cell phone battery is that good, but it appears to be.

I don't know if any heart rhythm analysis can be done with audio. As far as I know it's all ECG (electrocardiogram, or EKG if you use the German spelling). But if you got the electrodes to deliver energy to the chest, you already have the wiring needed for ECG and it's attached to the patient. So that part doesn't require special hardware or action. Audio would actually require additional hardware (some kind of stethoscope).

There was a scene in a James Bond movie, I think in Casino Royale, where Bond is poisoned and he tries to defibrillate himself using some kind of mini-defibrillator that he has in the glove compartment of his car. That was a pretty small unit, I wonder if it was based on an actual product. Pretty sure it's not for sale in the U.S. at least.

What I can see is a kit like that, that you might carry in the car or backpack, in a first aid kit along with band-aids, gauze, tweezers and aspirin. It can plug into the smartphone for the battery and to analyze the heart rhythm (shocking a person with a beating heart can be fatal!). It could use the USB connection, provided the circuitry would allow a heavy current draw. That's a much more feasible solution.

Hey nerf_herder and everyone else in this blog,

you seem to have brilliant ideas and to know a lot about defibrillators and electrical engineering. I'm just now building a team and raising funds to develop the very low cost, but yet reliable and FDA-cleared smartphone AED every household can afford. Whoever wants to join the team, cooperate, advice, co-found ... is more than welcome. Just drop me an email at You guys are really needed as part of the team. Thanks!


If you're going to have a 911 drone, why would it wait to confirm that the emergency is medical before flying? For some police action, sending the drone to record video may be useful, either to encourage people fighting to break it up, or to determine exactly what happened to help the prosecution, and for a fire, flying around the house to try to identify which rooms have people in them might provide useful data.

Relatedly, when is calling 911 from a smartphone going to start automatically sending video to the 911 dispatcher, maybe even from the cameras on both sides of the phone?

While yes, most emergencies could benefit from having an emergency worker virtually present, I expect that there are a lot of 911 calls and many of them do not result in sending somebody out. So you probably do have to figure out if it's time to send a drone, because you only have so many drones and you want to prioritize.

Make the drones cheap though and you can send them off to the location on autonomous mode, and then the right person can take control when it gets there. Yes, even for a crime in progress, the idea of a police drone which is making video of you, and will follow you if you flee is a deterrent against crime, even better than cameras on all the street corners.

For fire, it is good to have drone to estimate the nature of the fire and how many trucks to send.

For car accidents it may not be of much value, though a picture of the accident geometry would be handy so they can clear the accident.

But you still want to prioritize for when a life is at risk if you only have a limited number of drones. But yes, if there are several sitting waiting for instructions, might as well dispatch to the calling location on auto. Worst case is it comes back and gets a battery pack swap.

I think drones equipped with modern equipment would facilitate more intervention work crews and would save many lifes.

For police use it can also carry some pepper spray and a tazer. Maybe the defibrillator can do double duty on the later.

heart rhythm analysis can be done with audio. As far as I know it's all ECG (electrocardiogram, or EKG if you use the German spelling). But if you got the electrodes to deliver energy to the chest, you already have the wiring needed for ECG and it's attached to the patient. So that part doesn't require special hardware or action. Audio would actually require additional hardware (some

Cellphones are able to do everything needed. What we need is this: pads, wires to connect to a power supply big enough to supply ONE SHOCK. Why one? because that's the one that usualy resets the heart if you can get to the patient soon enough, and it will enable a device cheap and small enough for anyone to be willing to carry it around. consequently, there will be more people carrying them, perhaps even more than one at any incident of cardiac standstill. make it connectable with usb cables; the big hurdle is getting people who are in panic mode able to do what's necessary. The phone will do everything else. Attach the pads to the patient (which are already attached to the power gadget), plug usb to phone and gadget. the phone will know the gadget is attached and activate the app, whch will automatically read the rhythm and tell the rescuer what to do. It can even be made to voice activate a shock so after attaching the pads, no one has to ever touch the patient, and since it only gives one shock, there's little danger of rescuers being shocked. Cost: pads, wires and one shock power supply; maybe $9.99? App downloadable for free, and everyone already has usb cables (except for stupid apple, of course, but something like this should be enough to get them on board with what the rest of the world uses for cell phone connection). Hell, for that price, you could make them as stocking stuffers for chirstmas.


Add new comment

Subscribe to Comments for "A drone with a tele-doctor and defibrillator in 100 seconds?"