A survey of 911 dispatchers reveals the cost of bad technology
washingtonpost.com
washingtonpost.com
Its not just 911 calls from cell phones, its the whole stack. This will hopefully illustrate the problems we have;
Last night we were dispatched to an MVA (accident) with injuries. The 911 call came in from a passer by on a cell phone, but gave the wrong address.
A police officer near by, responded as well. When the officer arrived, he told Dispatch that the accident was on XYZ st. Dispatch corrected our address on our MDT (the computer in the Engine) and updated over the radio. We drove to the intersection, but couldn't find the accident.
I am the driver of the engine, so I ask my Captain if we should just head south a litter further to gain some elevation and look for lights. This is all while making appropriate radio transmissions. We drive a little further south up a small hill and find the Accident. The accident is on the street the officer said it was on. Dispatch didn't have an updated map that reflected the connection of two streets and the streets name change. They had dispatched us to the old intersection.
We were delayed about 4 minutes and the injuries were not serious, thankfully. However, this is pretty common.
We have two laptops on every truck. We have a brand new digital radio system. We have a year old CAD (dispatch) software package. We use electronic patient reporting and transmit full telemetries (patient reports) via a shitty web app.
Our laptops do not cross pollenate (share data). Our radio system doesn't work with police unless we coordinate beforehand or activate the MMRS system which is reserved for homeland security type events. The radio system has all kinds of other problems, it is fully software defined radio. Our patient data from the EKG monitors do not talk to the EPCR (electronic patient...) software. Our EPCR data does not cross pollenate with the shitty web app for transmission to the hospitals. CAD; Maps are a cluster f* and its too expensive to use Google maps apparently. The CAD software has all kinds of bugs for data extraction Ie. calls are doubled, missing and more.
Believe it or not we have Fire/EMS departments from around the country visiting us as we are very, very advanced in terms of technology in the Fire/EMS service. I've made two android apps for our department to use, an easy SOP reader and a Fire pump calculator. I am actively trying to come up with solutions to fix our problems via MORE! software...haha
In the case of the former, I wonder if OpenStreetMap is any better than CAD Maps' existing map source.
This doesn't come with the maps data itself, though.
That's just fundamental. It's not surveillance state stuff.
At one time, this problem was limited to a smaller population of cellphone users. Most people in one of these horrible car accident situations, etc didn't have access to a cellphone, so they just died.
Now, landlines are a thing of the past, so calls from office buildings, homes, etc are coming in without actionable information.
Speaking on the disappearance of land-lines, another issue is the increase of VoIP subscribers. Keeping address information up-to-date is the responsibility of the subscriber. If they take their VoIP service to the cottage, and call 911, emergency service could be dispatched to their home residence instead.
That may be, but not being able to locate people in an emergency is a legitimate problem, and it's happening because people are ditching their landlines (which did have known locations).
I live on the shore of Puget Sound in Washington State, right by the water. This means my cell phone is much more likely to connect with a cell tower on the opposite shore -- and people on the opposite bank get connected to cell towers on this side. This means depending on a cell tower for location is totally misleading for all these people. The only people who can be automatically located are those who (a) have smartphones with built-in GPS, (b) the GPS location is set to be included with each call, and (c) are outdoors when they make the call.
It's a mess, and conspiracy theories aside, it needs to be fixed.
To get that kind of data from your OS to a 911 dispatcher simply by placing a call would be to concede the ability for that kind of data to be leaked over telephony at all, or otherwise constantly uploaded to telcos/authorities.
Rather than a single number to call, show me the resources available in my local area, in realtime -- distances to a level I/II trauma facility, hazmat, fire, police, tactical, etc. (some of this might be restricted for security reasons, but "infinite" resources like a police hq or hospital wouldn't be).
I specifically want to be able to report a traffic accident in e.g. Pleasanton and get connected directly to CHP, rather than going through local PD first.
Voice is great, but I'd also like to get updates on ETA, and the ability to send more accurate information to the first responders than goes through 911 dispatchers -- specifically, the person on scene should see exactly what is being sent to the first responders to check it for accuracy. (maybe not so much in security/police, but in fire/medical) Sending photos and essentially a map could be really helpful.
The easiest commercial market is probably regulated-industry safety professionals (power plants, chemical plants, etc.) which have unique needs. One option might be deploying it within radius of these facilities to the general public so they can get updates on things like gas releases or accidents.
The longer the travel time between first responders and victims, the more useful this would be. In Palo Alto, you're going to get police/fire/medical quite quickly. In a poor city like Detroit, calls get triaged, and in actually remote areas, even more so -- yet, people often do have smartphones and at least 3G now.
It would also be a big win in markets where private ambulances and other options exist (i.e. most of Asia); knowing that you can get a private ambulance for $30, or take a taxi for $10, or wait 2h for state services, would be really useful.
Essentially drchrono or practice fusion for EMS/fire/police/etc -- less featureful and customizable than the existing solutions, but vastly easier to deploy.
I wonder if it would be a good customer development/market research strategy to volunteer with some local departments (EMT-B certification is pretty easy; it's a fair amount of time to volunteer though). Might be a good place to hire, as well.
Smartphones should have an emergency button. Pressing the button should automatically call 911 (or 112) and speaking using a Text-To-Speech TTS software (Siri, etc.) the name of the phone owner and the GPS coordinate, street address, etc.
This technology is already built in e.g. Ford cars (using Nuance/Microsoft speech technology, using automatic Bluetooth to call via the phone) - at the same time as the airbag goes off the car calls the emergency number and speaks.
And there are a number of (smart)phones that have an inbuilt emergency button that calls or SMS a predefined (but editable) number (like relatives).
Seriously, it it kind of scares the crap out of me visualizing how I'd dial for an emergency if shit hit the fan and for whatever reason I wasn't my normal self.
> Smartphones should have an emergency button.
There is an emergency call button on the iPhone lock screen, because you have to be able to place such calls even if it's not your phone, or it's not activated, or the SIM is not in, etc.
> Pressing the button should automatically call 911 (or 112)
Actually the same thing, since dialing the emergency number doesn't just call a number, but starts a GSM emergency call. That's why dialing 911 on an American phone works even if you've taken it with you to another country. (This doesn't quite work in countries with separate police/fire/ambulance numbers.)
Actually, if you tell Siri to "Call 911" I believe it just does it with no confirmation, unlike other numbers. Not sure what information is provided by cell phones on these calls, though.
Imagine you have a stroke, or your hands are full of blood, or you fall and brake your hands ...then you cannot use your touch screen smartphone and/or speak comprehensible words!
In such emergency situations a text-to-speech would save life's. You may be able to press a physical button or the G-sensor detects the accident and reacts automatically.
Just don't ever drop your phone.
This would be so easy to implement, can it even really be said to be a technology problem?
For those PSAPs that do have NG911 equipment, managing an SMS "call" is tricky because SMS messages are not sent on a persistent connection. Each message requires routing through the wireless network to a selective router which then routes the call to a PSAP, and the PSAP routes the message to a call taker. So somehow subsequent messages need to be able to find their way to the same call taker that handled the first message.
- How do all of the routing points know what "call" each message belongs to?
- How do they know when the call ends?
- What happens when the caller's location changes and the wireless network routes through a new selective router?
- What if a PSAP is busy or offline? Unlike normal SMS, messages need to be re-routed.
- 911 calls need to be recorded. Who is responsible for storing the SMS conversations? Sometimes PSAPs are required by a state to record conversations. If the SMS solution is a web-hosted application, state laws may need amending to allow the application provider to record the SMS conversation.
The wireless network, selective router, and PSAP equipment may each be managed by a different entity, so it requires new laws, standards and equipment to coordinate a 911 SMS conversation correctly compared to normal SMS calls where the number that is being sent to maps to a single subscriber.
My guess is they're being volume-limited to protect people's hearing, but in a high ambient-noise environment (like, standing beside an interstate) it really needs to be louder.
In some cases the caller is not able to communicate verbally because they're choking, hiding in the closet from an intruder, or any number of other reasons. In those situations, being able to reach 911 by SMS would allow them to give a location. You can't yet, but it's being worked on.
But yeah, if you're in a city with huge buildings blocking GPS signals and 50 floors of little apartments stacked on top of each other, then a rough street address may not help enough anyway.
First, cell towers do better than mere signal strength triangulation -- they use signal arrival time and round-trip time to get a pretty good fix on a phone's location. Not nearly as good as GPS, but reasonably good.
Second, smartphones have GPS receivers, and some of those allow the GPS position to accompany the outgoing call. That's much better, but it sometimes requires that the call be placed outdoors.
I think in the future, all cell phones will have GPS, and a few other improvements will come along to allow much better positioning.
Do you have a source for this? As far as I know this is not true. They use signal strength and knowledge of cell tower antenna locations, but not time of flight of the radio signal, as GPS does.
http://en.wikipedia.org/wiki/Mobile_phone_tracking
Quote: "Using the SIM in GSM and UMTS handsets, it is possible to obtain raw radio measurements from the handset.[6][7] The measurements that are available can include the serving Cell ID, round-trip time, and signal strength. "
http://wrongfulconvictionsblog.org/2012/06/01/cell-tower-tri...
Quote: "Within each sector, the tower can make a measurement of how far away the transmitting cell phone is. This is done by measuring signal strength and the round-trip signal time."
Neither of the above references is very specific, but I recently read a paper that went into great detail, and showed graphs of the areas with the greatest accuracy, all based on arrival time measurements. Some of it may have been speculative based on future improvements in the system. But I can't locate that specific paper.
Here's another reference:
http://www.cnet.com/news/feds-push-for-tracking-cell-phones/
Quote: "T-Mobile, for instance, uses a GSM technology called Uplink Time Difference of Arrival, or U-TDOA, which calculates a position based on precisely how long it takes signals to reach towers. A company called TruePosition, which provides U-TDOA services to T-Mobile, boasts of "accuracy to under 50 meters" that's available "for start-of-call, midcall, or when idle.""
You know, now that I've read these articles, it sounds more like a future technology than something that's in place right now, except for a few networks.
http://www.trueposition.com/about-trueposition/resource-cent...
Accuracy of 50 meters is a bit disappointing, but hopefully the technology has room to improve. It's also disappointing that they're focused on the benefits for law enforcement and carriers rather than handset makers and end users.
That's just politics. The cell carriers know they need to get regulatory approval from government agencies before they begin to sell the technology to the public. That regulatory approval is more likely if they can show a governmental advantage, like pinning down the location of a drug dealer or a terrorist chatting on his cell phone.
The real advantage for the carriers, by the way, is they can make a planned tower switch for a moving phone more accurately if they have both position and velocity information. It's hard to get a velocity vector if you can't resolve a position to better than 3/4 mile, what you get with signal strength measurements and no time-of-flight. (The Doppler effect could provide velocity as well, but without a very good direction of motion.)
Then the feature would be useful in day to day life and would be well supported. We have to stop treating emergency communications as something special. The actual problem here is a lack of proper standards.
You can't easily query those location details quickly.
Something to think about.
How can the cell companies share data with the NSA against the wishes of their customers, yet not share information with 911 which is clearly in the best interests of the customer?
I just don't get it. Do corporations really don't give a shit about whether people live and die?
Honestly I'm enthused to hear the initiative, critical thinking, and passion behind your ideas here. I personally agree with many of you that the first problem here IS NOT technology. It isn't that we don't have the technology, it is that the technology isn't being properly applied nor has been there been a willingness to adopt it. A big reason for this is that the people making the decisions of what to use and provide (telcos and CAD companies) are not the people who face the consequences when response arrives too slowly (regular people like you and me).
As monkmartinez said, this is a full-stack issue. Here's the full workflow assuming what is state-of-the-art in a 911 center.
You need help. Bring out your smartphone and press 'Emergency' if you have iOS. Tap '9' '1' '1' and dial.
Call rings once, maybe twice. Depending on the area, this could be a long time.
Call is routed to nearest PSAP
They pick up and say '9-1-1 emergency.'
You say something about what's going on.
Their first question, "Where are you located?"
Most often, you don't exactly know. so if you were in SF, "I'm in SOMA, near the ballpark."
They want more detail, cause they want to avoid what happened with monkmartinez below. "Sir, can you name cross-streets?"
Hopefully you know. "3rd and brannan". Now they have something they can use to dispatch responders.
If you don't know, they will keep asking for more specificity. At the same time, they are using their console (which looks like it was built in Fortran) to manually refresh to try and get your coordinates. This takes anywhere from 45 seconds to 6 minutes, the whole time you have to be on the call without hanging up.
Let's say they get the coordinates. The accuracy can be anywhere from 150 meters to 4000meters. Also, they come in as RAW coordinates (37.7756, -122.4193).
They then write these coordinates on paper, go to Google Maps on a separate console and type them in. THEN they have an address that they can dispatch responders to. It is insane.
There is so much room to improve this. 9-1-1 is a public service that people depend as part of the social fabric that enables our lives. It is crazy to me that this is such a problem, and I know that it is one that has to be fixed. We think that fixing this will definitively make the world a better place and enable many people to live better lives.
If you want to know more about our company and what we are working on: angel.co/bluelight-1 or email preet@getbluelight.com
Edit: Another issue is the people making the purchasing decisions on the public safety side do NOT understand technology. The men and women who are Chiefs or Directors were not born into technology. They are "old school" firefighters/paramedics and with that mindset comes a lot of baggage. Most of them are downright scared of technology. This is rapidly changing as the demographics of the fire/ems service are in total flux. Baby boomers are retiring because the job is physically demanding and your body wears out. Newer guys/gals like myself (9 years) embrace tech and are active to find ways to leverage it.
I have lots of ideas that I am working on for Android and the fire/ems service, but are mostly related to education, reference, promotional exams and such Ie. one man projects. These big problems are complex beasts that would require me to quit driving a fire engine (sometimes, I can't believe I am actually paid to do that) while improving my programming ability full time to gain employment in the software industry... its a conundrum that I've wrestled with for some time.