06 Sep

Google Car will be Life Changing

Google Car

Technology progresses in geometric speed.  With every passing day, technological breakthroughs keep arriving faster.  What more, each change has a much bigger impact than the previous change.  Ray Kurzweil predicts that at this rate, technology would’ve progressed sufficiently for man to merge with machines.  He has named this incident the “technological singularity“.  Because our brains are not equipped to perceive exponential curve, we find it extremely difficult to wrap it around the accelerating power of technology (TED video).

Google Car is a strong reminder to the skeptics that rate and magnitude of technological changes do accelerate.  Though Google Car per se is of only peripheral interest to this blog, let us take a closer look at it to help us comprehend the speeding up of technological changes.

Born out of a 2005 DARPA challenge tackled jointly by Stanford and Google, Google’s driverless car (video) pet project has taken a life of its own: The cars have driven more than 500,000km  accident and driver free, three states of US have permitted plying these cars on their roads and many more states and countries are to follow suit.  What looks like a pet project of a billionaire pair and a bunch of geeks is anything but a toy.  This innovation is about to make big changes to the way we live, and the change is going to happen much more quickly than we can imagine.

90% less accidents

The most emotionally appealing advantage from an self-driving car technology is this: 90% of the traffic accidents are caused by human error!  In the US alone, Google expects that its driverless technology will save 30,000 lives an year, avoid additional 2 million injuries and reduce accident related expenses by at least $400 billion.

Save commute time

Because driverless cars will be networked and exchange traffic information, drive with a much lesser physical clearance from each other, avoid collisions more effectively, etc., driverless cars can run faster, more cars can be accommodated in the given road space and they effortlessly select most optimal routes.  Hence, driverless cars will substantially reduce traffic jam and reduce commute time, to 90% by some rough estimates.  A 2012 study estimates that traffic congestion cost about $100b in US alone!

Less Cars on the Road

A car is one of the largest personal expenditure met today, still a car sits underutilized for most of the day.  This is with driverless cars.  In the era of driverless vehicles, most of us will seek to reduce under-utilization of assets by using common resources.  These resources will be operated by businesses that appear like cab companies.  The will have cars parked all over the city, in designated locations.  People hail them electronically, hop in and hop off.  Because the human driver is eliminated, driverless taxi service will be available at just a fraction of the cost we pay our cabs now (which in itself is much smaller than what we usually pay to self-owned car).  If the trend catches up, car pooling will automatically evolve out of this and the cost of using a car will sharply plummet.  Along with this, the number of cars will come down too.  By an optimistic estimate, driverless cars may reduce the total number of cars on the road by as much as 90%!

With 90% of the cars, traffic jams, commute time and traffic accidents all will further reduce!

What changes will happen

Driverless cars will make professional drivers obsolete.  US is already talking about truck drivers being eliminated over the next few years.  The trend will continue and accelerate.  Along with drivers, driving schools will eventually become redundant too.

With the cars being controlled and guided by pre-programmed rules, traffic violations will vanish.  With that traffic policing will shrink and take new roles.

Substantially reduced accidents will eventually put emergency medical services out of business.  Along with that, revenue of car insurance companies, and the companies that offer road accident covers, will shrink.

With the utilization of the cars increasing from the present levels, need for parking space and facilities will reduce.  With substantial reduction in the number of cars, and traffic snarls, construction of new roads will slow down and consumption of fuel plummet (on the other hand, if we discover cheaper sources of energy, car traffic on the highways will increase).  With better utilization of resources, cost of using a car per mile will plummet.

With number of cars reducing (after an initial hike owing to people converting to driverless), many car companies will go out of business.

Cab companies will flourish, but only if they rediscover themselves ground up.  With no driver inside a car, in-car entertainment will take new dimensions.  We will see the birth of a new vertical called ‘car-apps’ and the app makers coming up with several innovative solutions that we haven’t even felt the need for today.  Google, of course, is expected to share the driverless OS (like it shared the Android OS of the mobile phones) and build another mammoth revenue stream out of it.

So much for a pet project.  The funniest thing is, the technological innovations that gave us driverless cars is trivial comparing with, say, coming up with an intelligent search engine like Google (we easily forget that what a fine piece of peerless technology the Google search is).

How soon

In 2008, self driving cars could drive two blocks of closed loop at 40kmph.  By 2012, driverless cars have already driven 500,000km  in real life traffic.  With that kind of driving experience, Google must have pretty much figured out the driving algorithm.

A driverless car must have four additional feature

  • Machine operated controls (trivial challenge)
  • Central silicon brain (trivial challenge)
  • Driverless OS (Google already has it and it shouldn’t be too difficult for anyone who sets out to build it.  Given that it is just a piece software with few million of dollars of investment, the bar is not high at all)
  • Sensors, consisting of RADAR, gyro, GPS, etc. (existing technology)

As of today, the cost of the sensors runs into hundreds of thousands of dollars and they need to go through further miniaturization.  This is the only piece of technology that is keeping the driverless car from becoming a mass market product.  There is nothing fundamentally revolutionary about any of the technologies that make driverless cars possible.

If we take a closer look at the sensors, we will realize that

  • Microwave generator in kitchen oven costs just a few dollars; other than this, everything else is just solid state signal processing
  • Gyroscope (usually, a fiber optic gyro), again, is largely software
  • GPS has already become dirt cheap
  • The cost of putting them all together, as of today, is high.  But that is because only few small companies with limited resources have been developing these technologies for the tiny industry of land mapping.  When mass market players jump in, cost of these sensors will plummet.

With time, these sensors will cost just a little more than a microwave oven.  We shouldn’t be surprised if inexpensive driverless cars become available even before all the countries in the world have permissions in place for plying a driverless car!  In less than three years, driverless cars will become common scene in select places on the planet.  In about 10 years, driverless cars will be a common thing anywhere in the world.  That should convince the skeptics that the Singularity is Near (book)!

Note 1: Funny, the Google Chrome browser I am using to write this article keeps reminding me that “driverless” is not a valid word

Note 2: This post is largely based on the book Driverless Cars: Trillions are up for Grabs

Leave a Reply

Your email address will not be published. Required fields are marked *