Self-driving cars projected to reduce injuries by 90 percent, save $450 billion annually

By on October 24, 2013, 7:00 PM
google, autonomous cars, self driving cars

Driver error is the number one cause of automobile crashes so what would happen if you removed humans from the equation? According to independent research by the Eno Center for Transportation, vehicle-related injuries would drop by 90 percent and save the US economy roughly $450 billion each year.

The group discovered that 40 percent of fatal crashes in the US involved alcohol, drugs, fatigue or distraction – all metrics that wouldn’t affect an autonomous vehicle. Even in cases where a vehicle is primarily responsible for an accident, human elements like not paying attention and speeding often contributed to the occurrence of crashes and / or the severity of injuries.

The adoption rate of self-driving vehicles among consumers will of course play a big role in how many accidents can be avoided and how much money the economy could save. For example, if one in every 10 car was replaced with an autonomous vehicle, it would reduce crashes and subsequent injuries by roughly half and save around $25 billion each year.

To realize 90 percent safer roads as mentioned in the introduction, an equal 90 percent adoption rate would be necessary while hitting the $450 billion savings mark would require nearly 100 percent adoption.

Several self-driving cars are in the works from the likes of Google, Mercedes, Nissan and Toyota but thus far, the high cost of necessary components and lack of general regulations have only helped to slow down progress. It’ll eventually happen but it’s just a matter of how long it will take.




User Comments: 36

Got something to say? Post a comment
cliffordcooley cliffordcooley, TechSpot Paladin, said:

As long as they stayed between the lines, they would be safer than someone texting and driving.

2 people like this | cmbjive said:

The problem with projections is that they are projections.

2 people like this | Burty117 Burty117, TechSpot Chancellor, said:

So who's fault is it when one of these computer controlled vehicles crashes into someone else? Or runs someone over? Or simply fails in some way?

No No, I don't think so, What should be happening is making the test to pass harder and the theory harder, they should spend this money on really drumming in the cold hard truth that you CANNOT text and drive or use your phone and drive unless you have Bluetooth built in, and even then it should really only be used if needed not entire journeys.

This just spells trouble in my eyes, sensors can get damaged or covered in mud, electronics DO fail, in fact, Electronics in cars are notorious for failing before anything mechanical does, if the electronics controls the mechanics? I cannot see this reducing anything other than some extremely strange court filings when families have no idea who to sue since it's a computers fault their little girl got ran over.

psycros psycros said:

Implementing and maintaining the infrastructure to make this work would cost far more than $450 billion a year. For a lot less money you could make people retake their driver's test when they renew their license every four years, hire more traffic cops, etc. Crazy, I know.

davislane1 davislane1 said:

And the savings is...1.6% of annual GDP. Given the U.S. govt's impeccable foresight into such matters, the consequent reduction in revenue from gas taxes, "death" taxes, medical taxes, and income taxes, would necessarily lead to said savings being erased by an increase in existing taxes or the creation of a new one, if not both.

Now, economics aside, I'm all for people adopting self-driving cars. I prefer to drive manually and anything that eliminates congestion is a winner in my book.

davislane1 davislane1 said:

Implementing and maintaining the infrastructure to make this work would cost far more than $450 billion a year. For a lot less money you could make people retake their driver's test when they renew their license every four years, hire more traffic cops, etc. Crazy, I know.

Stronger licensing requirements is what we need. If driver error is the leading cause of accidents, fix the driver not the car. As for adding more cops... No. Not unless they're all sitting exclusively in school zones and intersections.

Guest said:

Bring on the self driving cars. Only problem is that they'll be prohibitively expensive for at least 15 years but god I would love to own one. I drive 500+ miles a week, I'd love to be able to do something productive with that time. I stare out a windshield so much.

Dear Google, my company will test a fleet of cars for you, free of charge.

m4a4 m4a4 said:

Remove the human (*cough*stupidity*cough*) element out of driving? Sounds fine to me (at least when they system is acceptable). Too many ****** on the road anyways...

Edit: I didn't know id10t was considered a bad word -_-

davislane1 davislane1 said:

Edit: I didn't know id10t was considered a bad word -_-

HR says it's hurtful.

Guest said:

Consider this:

If automated cars actually work, then there shouldn't be crashes. If that's the case, then we shouldn't need car insurance, or our premiums should be nearly nothing because of the low accident rate.

If we need car insurance, then we're saying automated cars don't work, are infalliable, and potentially dangerous. In which case, why should they be on the road.

So which is it?

drjekel_mrhyde drjekel_mrhyde said:

Well I guess Mayhem will have a permanent on Law and Order SVU

3 people like this | JC713 JC713 said:

I cant trust self driving cars until they are secure and cannot be hacked.

They need quantum encryption!

Guest said:

This would be great, there is no real infrastructure needed to make driver less cars a reality. Each car needs adequate sensors and programming to react to immediate threats. And for navigation each car could just use GPS systems to find ways around. I'm sure advanced sensors could keep cars in their correct lanes of traffic. Of course each car can switch to manual operation in circumstances where GPS is not available or other issues arise.

Guest said:

Computers can never make predictions of dangerous situation. Two kids playing with a ball beside the road are just irrelevant objects to this computer, but a human might adjust his speed, just like he would do based on experience with the road, who live there, warning signs, the weight of his load, etc. We can also anticipate on drivers, based on their behavior, age, whether we had eye contact, the state of their car, the state of their load, etc.

The Eno Center for Transportation also doesn't take into account non-American roads, which are often way more complicated to drive. Take the Google car and plant it cities like Amsterdam, Athens, Paris, Madeira, Istanbul, etc.... completely different world.

Etc, etc, etc.

Those numbers are just accurate as waving your finger in the air to measure what the weather is going to be tomorrow.

Scshadow said:

So who's fault is it when one of these computer controlled vehicles crashes into someone else? Or runs someone over? Or simply fails in some way?

Actually, a potential precedent to the legality of the self-driving car actually exists. Inevitably the fault is with the manufacturer. Much like with vaccine manufacturers. See vaccine manufacturers are protected from liability to an extent because the government considers the benefit too great. As long as vaccine manufacturers aren't negligent to the due diligence of testing a product and complying with regulations, they aren't held liable. The same could potentially apply to fully autonomous vehicles. The benefit is too great that we can't have fear of liability completely stop autonomous vehicle development in its tracks. Once the same precedent is applied to manufacturers, the insurance companies would need to be forced to apply some sort of no-fault insurance coverage policy. We need some kind of "for the greater good" policy to make autonomous vehicles insurance coverage possible.

Consider this:

If automated cars actually work, then there shouldn't be crashes. If that's the case, then we shouldn't need car insurance, or our premiums should be nearly nothing because of the low accident rate.

If we need car insurance, then we're saying automated cars don't work, are infalliable, and potentially dangerous. In which case, why should they be on the road.

So which is it?

All cars break. You're saying that autonomous vehicles would have to have breaks that never fail, engines never blow, steering columns never break, tires that never blow before you'd consider them street worthy? I'm sorry your short-sightedness cannot see the vast vast benefit of having autonomous vehicles. As for logic errors, they aren't likely. The redundant sensors in current Autonomous vehicle design and the vigorous testing that entities like Google are doing, they are all to insure that its extremely unlikely that these machines will be anything less then extremely reliable logic-wise.

thewind said:

The people who think this is dumb are the stupid ones! 1st if you have multiple safe profs such as if a sensor is not responding the car slows down to a stop and alerts the driver to take over! not hard! Duh! also google has done over 1,000,000 miles of testing!!!!!!!! of which only 1 car was in an accident and it was the other guys fault! It don't take a genius to know if you've driven that much you probably been in more than 1 accident! I myself have been in 12 accidents some my fault, most weren't! and I'm only 26! 2nd people say electronics fail! I have a 30 year old calculator that still knows 2+2=4! Point being is that people make electronics fail. Yes once in a while things do fail due to a compositor failing or something but theres things you can do to make those last even longer... So lets say the average car life is 18 years. Make sure the parts in the computer can last 30 and then have mandatory to be replaced at 20 years! IF a person has there car that long then can just have the parts updated! There's so simple answers for things. People try to make things way to complicated! and +1 for Scshadow! Loved the comment!

Guest said:

What's the projected cost of 100% upgradation to the autonomous vehicle platform?

What's the associated cost of 'fixing' all the traffic signals, maps etc perfectly and regularly maintaining them?

Scshadow said:

What's the projected cost of 100% upgradation to the autonomous vehicle platform?

What's the associated cost of 'fixing' all the traffic signals, maps etc perfectly and regularly maintaining them?

The LIDAR is going to be the single most sophisticated and expensive device in the fully autonomous vehicle, but we really don't know for sure what it will cost by the time it enters mass production. My research didn't really go that far but I know the LIDAR currently used in google's vehicles are around $70,000 alone. Needless to say, the dependability of these Sensors has to be of a grade far higher then what anyone knows in our day to day lives.

As for "fixing" our roads, we don't. They're ready. GPS maps are merely a guideline to actual navigating which will be calculated in real time, with real time readings from the LIDAR radar. The traffic lights aren't even close to being a problem. One of the few problems left to work the kinks out is properly reading hand signals if a human being is redirecting traffic but that was a year ago at least. Pretty sure that would have been refined by now. By comparison regular traffic lights are absolutely nothing to interpret correctly.

Wendig0 Wendig0, TechSpot Paladin, said:

So, will I still need a designated driver when I go to bars/clubs? Will I still get a DUI if my car is doing the driving?

Guest said:

I think a self driving car would be great for me. I am too impatient. I'm too confident and end up speeding and having road rage. While in 20 years driving I have never had an accident (apart from my car bottoming out during a snow storm), I have received many speeding tickets. Less speeding tickets, and less stress would be a good thing. Plus I could get things done while my car drives me to where I want to go. Definitely a win win for me.

Guest said:

My personal experience/observations contradicts the theory that electronics fails first. Electronic failures are expensive to fix, therefore you may remember them better, but they are less frequent than 10% of all failures.

1 person liked this | Burty117 Burty117, TechSpot Chancellor, said:

All cars break. You're saying that autonomous vehicles would have to have breaks that never fail, engines never blow, steering columns never break, tires that never blow before you'd consider them street worthy? I'm sorry your short-sightedness cannot see the vast vast benefit of having autonomous vehicles. As for logic errors, they aren't likely. The redundant sensors in current Autonomous vehicle design and the vigorous testing that entities like Google are doing, they are all to insure that its extremely unlikely that these machines will be anything less then extremely reliable logic-wise.

You replied with the above about Car Insurance, Car Insurance does NOT cover servicing and Brake pads or anything like that, Car Insurance covers you when you land in a ditch or run someone off the road or some else runs you off the road, Car Maintenance is purely down to the driver.

To everyone above me, Please note I do see the usefulness of Autonomous cars and how they COULD work, but the fact you need Insurance implies they expect these things to run people off the road.

Car Maintenance is also a factor, some people just don't look after their cars properly, they are going to fail and it will be the drivers fault.

I'd also like to add, have you seen how expensive car Services already are on cars filled with Electronics?! This will also drive up Servicing Prices Massively.

Edit: These Types of cars will work when they are fully Electric Powered and not Gasoline and all roads are fitted with equipment to help the car navigate. Until then, they are still dangerous.

H3llion H3llion, TechSpot Paladin, said:

Implementing and maintaining the infrastructure to make this work would cost far more than $450 billion a year. For a lot less money you could make people retake their driver's test when they renew their license every four years, hire more traffic cops, etc. Crazy, I know.

This is the future. Better implement such infrastructure earlier then later.

amstech amstech, TechSpot Enthusiast, said:

Maybe in good weather conditions. On sleet/snow/black ice etc covered roads they will be much more dangerous.

MilwaukeeMike said:

Save $450 billion? You know when we spend money, someone else makes money. Saving $450 billion sounds great, unless you're the owner of a tow-truck business or a body shop.

spydercanopus spydercanopus said:

Regulation has never sped the development of technology.

Guest said:

I don't think its going to be one or the other as in a car that is in total control or a human in total control. There is still going to be a need for humans to pay attention. The computer will help to keep a car in its lane and avoid accidents. The driver however will not just simply be able to take a nap in the back seat while driving across the country. They will still need to be monitoring the systems. Being a retired pilot I can tell you todays planes can fly to your destination and even land on their own bit a pilot is still very much needed and used for when conditions are anything less then perfect or when systems fail. There are a ton of fly by wire aircraft (computer controlled) in the skies today that just do not simply fall out of the sky.

1 person liked this | Scshadow said:

You replied with the above about Car Insurance, Car Insurance does NOT cover servicing and Brake pads or anything like that, Car Insurance covers you when you land in a ditch or run someone off the road or some else runs you off the road, Car Maintenance is purely down to the driver.

Read again, I never suggested that maintenance was covered by insurance. I'm offended by the accusation that I would be so naive and stupid to think that. My reference to brake failure is in the context of being a cause to an accident, obviously. The Guest implied that Autonomous Vehicles didn't belong on the road unless they were guaranteed 100% accident free. And because of mechanical failure (like brake system failure) and regardless of who or what is behind the wheel, the need for insurance is not a reflection on the reliability of these vehicles.

If we need car insurance, then we're saying automated cars don't work, are infalliable, and potentially dangerous. In which case, why should they be on the road.

As I believe I so aptly implied, this statement is a bad jump in logic.

mattfrompa mattfrompa said:

I will be more excited for self driving cars if it doesn't just become a "luxury" kind of feature. Imagine this; You work a full time job where public transportation is prohibitively expensive or unavailable and you also don't have convenient access to another driver. Now say you fancy an evening at the pub, well I think we would all agree it would be nice to be able to drink and still get home safely. Well with a self driving car AND a breathalyzer (obviously nothing new) you could require a test before activating "manual mode" not at all times, but perhaps based upon location or a schedule(say 7pm-6am). This would not only provide a convenient option for the driver, but protect others from stupid mistakes.(drunk driving)

cliffordcooley cliffordcooley, TechSpot Paladin, said:

Well with a self driving car AND a breathalyzer (obviously nothing new) you could require a test before activating "manual mode" not at all times, but perhaps based upon location or a schedule(say 7pm-6am).
Nah, thats where the smart car comes in, it will detect a driver swerving and take control until they present evidence they are not drunk.

2 people like this | wiyosaya said:

<...>

This just spells trouble in my eyes, sensors can get damaged or covered in mud, electronics DO fail, in fact, Electronics in cars are notorious for failing before anything mechanical does, if the electronics controls the mechanics? I cannot see this reducing anything other than some extremely strange court filings when families have no idea who to sue since it's a computers fault their little girl got ran over.

Yes, electronics do fail. This is one area where some sort of standardized reliability requirements would help, and as in aircraft, doubly or triply redundant systems would also be an advantage even though it will increase the price.

Sensors can also be monitored by the brains of the car, and the brains of the car could possibly detect when they are damaged, but not so much so that they no longer work.

As far as sensors getting covered in mud, there are two solutions to that. 1. Put them in an area of the vehicle that is not as likely to get covered in mud, e.g., the roof. 2. The electromagnetic spectrum is very wide, and some frequencies have a greater ability to penetrate materials since many materials have predictable absorption bands. Some regions of the spectrum may be able to see through a coating of mud. It is not all that difficult for some spectrum regions. For short distances, think mm airport scanners as an example of something that can see through a material that is not transparent in the optical region of the em spectrum.

Personally, I do not see that sensors are a problem if they are thought through and implemented with safety in mind, and I think safety (as well as convenience) is what any group researching autonomous cars has in mind.

More problematic is trying to get either people who think they can do anything they want or are in some way ignorant of the dangers to not text, to not use hand's free devices and the like. I don't care how much you educate anyone, there will still be people who are "too smart" or "too dumb" not to do stuff while driving that puts them and all others on the road with them in danger.

For me, taking the driver out of the equation is the best means. In addition, I think it would be wise to make autonomous vehicles talk to each other. That way, they could coordinate and cooperate between them to maintain safe driving distances. There is nothing like that now given that there are drivers out there who think they own the road and them getting to where they want to go as fast as they can is the most important thing to them.

Burty117 Burty117, TechSpot Chancellor, said:

Read again, I never suggested that maintenance was covered by insurance. I'm offended by the accusation that I would be so naive and stupid to think that. My reference to brake failure is in the context of being a cause to an accident, obviously. The Guest implied that Autonomous Vehicles didn't belong on the road unless they were guaranteed 100% accident free. And because of mechanical failure (like brake system failure) and regardless of who or what is behind the wheel, the need for insurance is not a reflection on the reliability of these vehicles.

As I believe I so aptly implied, this statement is a bad jump in logic.

I Read it again, It's below for reference:

Guest Said

Consider this:

If automated cars actually work, then there shouldn't be crashes. If that's the case, then we shouldn't need car insurance, or our premiums should be nearly nothing because of the low accident rate.

If we need car insurance, then we're saying automated cars don't work, are infalliable, and potentially dangerous. In which case, why should they be on the road.

So which is it?

And you replied:

All cars break. You're saying that autonomous vehicles would have to have breaks that never fail, engines never blow, steering columns never break, tires that never blow before you'd consider them street worthy? I'm sorry your short-sightedness cannot see the vast vast benefit of having autonomous vehicles. As for logic errors, they aren't likely. The redundant sensors in current Autonomous vehicle design and the vigorous testing that entities like Google are doing, they are all to insure that its extremely unlikely that these machines will be anything less then extremely reliable logic-wise.

Guest didn't mention anything about breaking down, he said work, if they worked like intended then why do we need insurance? That is what he was querying, you went and replied about cars breaking down, that implies it worked as intended in the first place.

Anyway don't get me wrong, I bet they have tested them to bits, just like lots of stuff today is tested to death (literally) but fact of the matter is these cars will NOT work in places like the UK, or even Europe as a whole. They just won't, our roads are not only extremely old but also very difficult at times, These silly autonomous cars will either drive incredibly slowly (like max 20mph) or will be going down the wrong way down one way systems, getting confused at massive roundabouts and so forth.

In America, sure, why not, It's probably easier to install the right stuff on the roads for these cars and your roads are MUCH bigger, more room for error, but I'm sure if these went on sale, their would be plenty that sell, but I bet there are more people who actually enjoy driving and enjoy having control of a machine that can do over 100mph than a computer do it all for you.

Again I'm not completely against this, I know teenage girls over here that scare the living Jebus out of me how badly they drive, they are the kind of people who will benefit from this, but people like me who prefer to have control of my life when I'm going a decent speed and enjoy driving, no thanks, I would rather stick to good old fashion manual.

gamoniac said:

Regulation has never sped the development of technology.

And vice versa.

Guest said:

Do a search for the google self driving car. It works right now and as was said above the only real hold up is the cost of the sensor device. They will drive better in snow or ice better than 99% of the current people on the road. They can already detect and react to a child running out in the road and many other threats that you will never see due to the fact that they see in all directs at once and have a sub second reaction time.

I would expect to see this tech in some high end BMW within 10 years.

SNGX1275 SNGX1275, TS Forces Special, said:

If automated cars actually work, then there shouldn't be crashes. If that's the case, then we shouldn't need car insurance, or our premiums should be nearly nothing because of the low accident rate.

Yep, follow the money. With technology advancements, someone is going to lose business. If that someone has enough money/power, there will be some resistance (political campaign donations). Another thing along the line of insurance, is think of all the money involved in DUIs. If a car can prevent DUI accidents, suddenly that market goes away (DUI lawyers, county court income, community service hours, whatever else). Lots of consequences to having an accident free environment...

cliffordcooley cliffordcooley, TechSpot Paladin, said:

community service hours
Damn and our ditches have been so much cleaner, since people have learned how difficult it is to actually clean them. LOL

Load all comments...

Add New Comment

TechSpot Members
Login or sign up for free,
it takes about 30 seconds.
You may also...
Get complete access to the TechSpot community. Join thousands of technology enthusiasts that contribute and share knowledge in our forum. Get a private inbox, upload your own photo gallery and more.